Coal mining induces changes in the nature of rock and soil bodies,as well as hydrogeological conditions,which can easily trigger the occurrence of geological disasters such as water inrush,movement of the coal seam ro...Coal mining induces changes in the nature of rock and soil bodies,as well as hydrogeological conditions,which can easily trigger the occurrence of geological disasters such as water inrush,movement of the coal seam roof and floor,and rock burst.Transparency in coal mine geological conditions provides technical support for intelligent coal mining and geological disaster prevention.In this sense,it is of great significance to address the requirements for informatizing coal mine geological conditions,dynamically adjust sensing parameters,and accurately identify disaster characteristics so as to prevent and control coal mine geological disasters.This paper examines the various action fields associated with geological disasters in mining faces and scrutinizes the types and sensing parameters of geological disasters resulting from coal seam mining.On this basis,it summarizes a distributed fiber-optic sensing technology framework for transparent geology in coal mines.Combined with the multi-field monitoring characteristics of the strain field,the temperature field,and the vibration field of distributed optical fiber sensing technology,parameters such as the strain increment ratio,the aquifer temperature gradient,and the acoustic wave amplitude are extracted as eigenvalues for identifying rock breaking,aquifer water level,and water cut range,and a multi-field sensing method is established for identifying the characteristics of mining-induced rock mass disasters.The development direction of transparent geology based on optical fiber sensing technology is proposed in terms of the aspects of sensing optical fiber structure for large deformation monitoring,identification accuracy of optical fiber acoustic signals,multi-parameter monitoring,and early warning methods.展开更多
Distributed fiber-optic sensing has become an indispensable tool for large-scale structural and environmentalmonitoring, where spectral interrogation of backscattering light enables high-precision quantitative measure...Distributed fiber-optic sensing has become an indispensable tool for large-scale structural and environmentalmonitoring, where spectral interrogation of backscattering light enables high-precision quantitative measurement ofexternal perturbations. Conventional spectral analysis methods, typically based on frequency-domain serialinterrogation or time-to-frequency mapping, face inherent trade-offs between measurement speed, dynamic strainmeasurement range, and system complexity. Here, we present a distributed frequency comb enabled spectrumcorrelationreflectometry as a universal spectral analysis framework that leverages optical frequency comb for parallelmulti-frequency interrogation, which is experimentally demonstrated in a phase-sensitive optical time-domainreflectometry (φ-OTDR) system. This method eliminates the need for large frequency scans, achieving more thantenfold improvement in measurement speed over the state-of-the-art spectral analysis methods. Compared to existingphase-demodulated φ-OTDR systems, this method enables vibration amplitude monitoring with a dynamic strainmeasurement range expanded by more than an order of magnitude, while intrinsically circumventing phaseunwrapping issues and interference fading. This work establishes a new paradigm for distributed spectral analysis,providing a flexible and robust platform for a wide range of sensing technologies, including Rayleigh and Brillouinbasedschemes, which may have significant impact for geophysics, seismology, civil engineering, and other fields.展开更多
A multi-stage stress relaxation test was performed on a granodiorite sample to understand the deformation process prior to the macroscopic failure of brittle rocks,as well as the transient response during stress relax...A multi-stage stress relaxation test was performed on a granodiorite sample to understand the deformation process prior to the macroscopic failure of brittle rocks,as well as the transient response during stress relaxation.Distributed optical fiber sensing was used to measure strains across the sample surface by helically wrapping the single-mode fiber around the cylindrical sample.Close agreement was observed between the circumferential strains obtained from the optical fibers and the extensometer.The reconstructed full-field strain contours show strain heterogeneity from the crack closure phase,and the strains in the later deformation phase are dominantly localized within the former high-strain zone.The Gini coefficient was used to quantify the degree of strain localization and shows an initial increase during the crack closure phase,a decrease during the linear elastic phase,and a subsequent increase during the post-yielding phase.This behavior corresponds to a process of initial localization from an imperfect boundary condition,homogenization,and eventual relocalization prior to the macroscopic failure of the sample.The transient strain rate decay during the stress relaxation phase was quantified using the p-value in the“Omori-like"power law function.A higher initial stress at the onset of relaxation results in a lower p-value,indicating a slower strain rate decay.As the sample approaches macroscopic failure,the lowest p-value shifts from the most damaged zone to adjacent areas,suggesting stress redistribution or crack propagation in deformed crystalline rocks under stress relaxation conditions.展开更多
Distributed optical fiber sensing(DOFS)technology has been widely applied in pipeline monitoring,seismic detection,and security protection due to its broad coverage,high sensitivity,and strong anti-interference capabi...Distributed optical fiber sensing(DOFS)technology has been widely applied in pipeline monitoring,seismic detection,and security protection due to its broad coverage,high sensitivity,and strong anti-interference capability.However,the acquired signals are typically noisy,exhibit complex temporal-spatial patterns,and contain high-dimensional categorical features,posing significant challenges for robust classification.To address these issues,this paper introduces an Inception-ResNet-based model for intrusion event recognition in DOFS systems.The Inception architecture extracts multi-scale features from complex vibration patterns,while the residual optimization of ResNet enables efficient deep feature propagation and stable training.Furthermore,to enhance model interpretability,a Grad-CAM-based mechanism is integrated to visualize class-discriminative regions in the vibration signals,revealing the patterns that most strongly influence the network's decisions.Extensive experiments demonstrate the effectiveness of the proposed approach,achieving an average classification accuracy of 92.6%,outperforming traditional deep learning networks even with significantly reduced training data.These results indicate that the interpretable Inception-ResNet framework not only accurately classifies complex one-dimensional sensing signals but also provides transparent and reliable support for practical DOFS applications.展开更多
To investigate the damage evolution caused by stress-driven and sub-critical crack propagation within the Beishan granite under multi-creep triaxial compressive conditions,the distributed optical fiber sensing and X-r...To investigate the damage evolution caused by stress-driven and sub-critical crack propagation within the Beishan granite under multi-creep triaxial compressive conditions,the distributed optical fiber sensing and X-ray computed tomography were combined to obtain the strain distribution over the sample surface and internal fractures of the samples.The Gini and skewness(G-S)coefficients were used to quantify strain localization during tests,where the Gini coefficient reflects the degree of clustering of elements with high strain values,i.e.,strain localization/delocalization.The strain localization-induced asymmetry of data distribution is quantified by the skewness coefficient.A precursor to granite failure is defined by the rapid and simultaneous increase of the G-S coefficients,which are calculated from strain increment,giving an earlier warning of failure by about 8%peak stress than those from absolute strain values.Moreover,the process of damage accumulation due to stress-driven crack propagation in Beishan granite is different at various confining pressures as the stress exceeds the crack initiation stress.Concretely,strain localization is continuous until brittle failure at higher confining pressure,while both strain localization and delocalization occur at lower confining pressure.Despite the different stress conditions,a similar statistical characteristic of strain localization during the creep stage is observed.The Gini coefficient increases,and the skewness coefficient decreases slightly as the creep stress is below 95%peak stress.When the accelerated strain localization begins,the Gini and skewness coefficients increase rapidly and simultaneously.展开更多
Defects in cast-in-situ piles have an adverse impact on load transfer at the pile‒soil interface and pile bearing capacity. In recent years, thermal integrity profiling (TIP) has been developed to measure temperature ...Defects in cast-in-situ piles have an adverse impact on load transfer at the pile‒soil interface and pile bearing capacity. In recent years, thermal integrity profiling (TIP) has been developed to measure temperature profiles of cast-in-situ piles, enabling the detection of structural defects or anomalies at the early stage of construction. However, using this integrity testing method to evaluate potential defects in cast-in-situ piles requires a comprehensive understanding of the mechanism of hydration heat transfer from piles to surrounding soils. In this study, small-scale model tests were conducted in laboratory to investigate the performance of TIP in detecting pile integrity. Fiber-optic distributed temperature sensing (DTS) technology was used to monitor detailed temperature variations along model piles in sand. Additionally, sensors were installed in sand to measure water content and matric suction. An interpretation method against available DTS-based thermal profiles was proposed to reveal the potential defective regions. It shows that the temperature difference between normal and defective piles is more obvious in wet sand. In addition, there is a critical zone of water migration in sand due to the water absorption behavior of cement and temperature transfer-induced water migration in the early-age concrete setting. These findings could provide important insight into the improvement of the TIP testing method for field applications.展开更多
Fiber-optic sensing technology has the advantages of passivity, anti-electromagnetic interference, longdistancemeasurement, high sensitivity and high accuracy, small size, and adaptability to harsh environments such a...Fiber-optic sensing technology has the advantages of passivity, anti-electromagnetic interference, longdistancemeasurement, high sensitivity and high accuracy, small size, and adaptability to harsh environments such ashigh-vacuum, high-pressure, and strong magnetic fields compared with the traditional electrical sensing technology.However, with the increasing application requirements, how to further improve the sensitivity of fiber-optic sensors,extend the detection limit and improve the maintenance-free capability has become one of the core issues of thecurrent research. This paper reviews the principle, preparation, and application of fiber-optic microstructured sensingbased on abrupt field type. It specifically outlines the development and applications of micro-nano optical fibers,photonic crystal optical fibers, optical fiber gratings and structured optical fibers, and lists the main preparationmethods of two types of micro-nano optical fibers from the basic theory of optical fiber microstructured sensordevices.展开更多
Fiber-optic distributed strain sensing(FO-DSS)has been successful in monitoring strain changes along horizontal wellbores in hydraulically fractured reservoirs.However,the mechanism driving the various FO-DSS response...Fiber-optic distributed strain sensing(FO-DSS)has been successful in monitoring strain changes along horizontal wellbores in hydraulically fractured reservoirs.However,the mechanism driving the various FO-DSS responses associated with near-wellbore hydraulic fracture properties is still unclear.To address this knowledge gap,we use coupled wellbore-reservoir-geomechanics simulations to study measured strain-change behavior and infer hydraulic fracture characteristics.The crossflow among fractures is captured through explicit modeling of the transient wellbore flow.In addition,local grid refinement is applied to accurately capture strain changes along the fiber.A Base Case model was designed with four fractures of varying properties,simulating strain change signals when the production well is shut-in for 10 d after 240 d of production and reopened for 2 d.Strain-pressure plots for different fracture clusters were used to gain insights into inferring fracture properties using DSS data.When comparing the model with and without the wellbore,distinct strain change signals were observed,emphasizing the importance of incorporating the wellbore in FO-DSS modeling.The effects of fracture spacing and matrix permeability on strain change signals were thoroughly investigated.The results of our numerical study can improve the understanding of the relation between DSS signals and fracture hydraulic properties,thus maximizing the value of the dataset for fracture diagnostics and characterization.展开更多
A distributed optical-fiber acoustic sensor is an acoustic sensor that uses the optical fiber itself as a photosensitive medium,and is based on Rayleigh backscattering in an optical fiber.The sensor is widely used in ...A distributed optical-fiber acoustic sensor is an acoustic sensor that uses the optical fiber itself as a photosensitive medium,and is based on Rayleigh backscattering in an optical fiber.The sensor is widely used in the safety monitoring of oil and gas pipelines,the classification of weak acoustic signals,defense,seismic prospecting,and other fields.In the field of seismic prospecting,distributed optical-fiber acoustic sensing(DAS)will gradually replace the use of the traditional geophone.The present paper mainly expounds the recent application of DAS,and summarizes recent research achievements of DAS in resource exploration,intrusion monitoring,pattern recognition,and other fields and various DAS system structures.It is found that the high-sensitivity and long-distance sensing capabilities of DAS play a role in the extensive monitoring applications of DAS in engineering.The future application and development of DAS technology are examined,with the hope of promoting the wider application of the DAS technology,which benefits engineering and society.展开更多
Phase-sensitive Optical Time-Domain Reflectometer(φ-OTDR)technology facilitates the real-time detection of vibration events along fiber optic cables by analyzing changes in Rayleigh scattering signals.This technology...Phase-sensitive Optical Time-Domain Reflectometer(φ-OTDR)technology facilitates the real-time detection of vibration events along fiber optic cables by analyzing changes in Rayleigh scattering signals.This technology is widely used in applications such as intrusion monitoring and structural health assessments.Traditional signal processing methods,such as Support Vector Machines(SVM)and K-Nearest Neighbors(KNN),have limitations in feature extraction and classification in complex environments.Conversely,a single deep learning model often struggles with capturing long time-series dependencies and mitigating noise interference.In this study,we propose a deep learning model that integrates Convolutional Neural Network(CNN),Long Short-Term Memory Network(LSTM),and Transformer modules,leveraging φ-OTDR technology for distributed fiber vibration sensing event recognition.The hybrid model combines the CNN's capability to extract local features,the LSTM's ability to model temporal dynamics,and the Transformer's proficiency in capturing global dependencies.This integration significantly enhances the accuracy and robustness of event recognition.In experiments involving six types of vibration events,the model consistently achieved a validation accuracy of 0.92,and maintained a validation loss of approximately 0.2,surpassing other models,such as TAM+BiLSTM and CNN+CBAM.The results indicate that the CNN+LSTM+Transformer model is highly effective in handling vibration signal classification tasks in complex scenarios,offering a promising new direction for the application of fiber optic vibration sensing technology.展开更多
Despite the extensive use of distributed fiber optic sensing(DFOS)in monitoring underground structures,its potential in detecting structural anomalies,such as cracks and cavities,is still not fully understood.To contr...Despite the extensive use of distributed fiber optic sensing(DFOS)in monitoring underground structures,its potential in detecting structural anomalies,such as cracks and cavities,is still not fully understood.To contribute to the identification of defects in underground structures,this study conducted a four-point bending test of a reinforced concrete(RC)beam and uniaxial loading tests of an RC specimen with local cavities.The experimental results revealed the disparity in DFOS strain spike profiles between these two structural anomalies.The effectiveness of DFOS in the quantification of crack opening displacement(COD)was also demonstrated,even in cases where perfect bonding was not achievable between the cable and structures.In addition,DFOS strain spikes observed in two diaphragm wall panels of a twin circular shaft were also reported.The most probable cause of those spikes was identified as the mechanical behavior associated with local concrete contamination.With the utilization of the strain profiles obtained from laboratory tests and field monitoring,three types of multi-classifiers,based on support vector machine(SVM),random forest(RF),and backpropagation neural network(BP),were employed to classify strain profiles,including crack-induced spikes,non-crack-induced spikes,and non-spike strain profiles.Among these classifiers,the SVM-based classifier exhibited superior performance in terms of accuracy and model robustness.This finding suggests that the SVM-based classifier holds promise as a potential solution for the automatic detection and classification of defects in underground structures during long-term monitoring.展开更多
The Carter model is used to characterize the dynamic behaviors of fracture growth and fracturing fluid leakoff.A thermo-fluid coupling temperature response forward model is built considering the fluid flow and heat tr...The Carter model is used to characterize the dynamic behaviors of fracture growth and fracturing fluid leakoff.A thermo-fluid coupling temperature response forward model is built considering the fluid flow and heat transfer in wellbore,fracture and reservoir.The influences of fracturing parameters and fracture parameters on the responses of distributed temperature sensing(DTS)are analyzed,and a diagnosis method of fracture parameters is presented based on the simulated annealing algorithm.A field case study is introduced to verify the model’s reliability.Typical V-shaped characteristics can be observed from the DTS responses in the multi-cluster fracturing process,with locations corresponding to the hydraulic fractures.The V-shape depth is shallower for a higher injection rate and longer fracturing and shut-in time.Also,the V-shape is wider for a higher fracture-surface leakoff coefficient,longer fracturing time and smaller fracture width.Additionally,the cooling effect near the wellbore continues to spread into the reservoir during the shut-in period,causing the DTS temperature to decrease instead of rise.Real-time monitoring and interpretation of DTS temperature data can help understand the fracture propagation during fracturing operation,so that immediate measures can be taken to improve the fracturing performance.展开更多
Curved geostructures,such as tunnels,are commonly encountered in geotechnical engineering and are critical to maintaining structural stability.Ensuring their proper performance through field monitoring during their se...Curved geostructures,such as tunnels,are commonly encountered in geotechnical engineering and are critical to maintaining structural stability.Ensuring their proper performance through field monitoring during their service life is essential for the overall functionality of geotechnical infrastructure.Distributed Brillouin sensing(DBS)is increasingly applied in geotechnical projects due to its ability to acquire spatially continuous strain and temperature distributions over distances of up to 150 km using a single optical fibre.However,limited by the complex operations of distributed optic fibre sensing(DFOS)sensors in curved structures,previous reports about exploiting DBS in geotechnical structural health monitoring(SHM)have mostly been focused on flat surfaces.The lack of suitable DFOS installation methods matched to the spatial characteristics of continuous monitoring is one of the major factors that hinder the further application of this technique in curved structures.This review paper starts with a brief introduction of the fundamental working principle of DBS and the inherent limitations of DBS being used on monitoring curved surfaces.Subsequently,the state-of-the-art installation methods of optical fibres in curved structures are reviewed and compared to address the most suitable scenario of each method and their advantages and disadvantages.The installation challenges of optical fibres that can highly affect measurement accuracy are also discussed in the paper.展开更多
Distributed acoustic sensing(DAS)is increasingly used in seismic exploration owing to its wide frequency range,dense sampling and real-time monitoring.DAS radiation patterns help to understand angle response of DAS re...Distributed acoustic sensing(DAS)is increasingly used in seismic exploration owing to its wide frequency range,dense sampling and real-time monitoring.DAS radiation patterns help to understand angle response of DAS records and improve the quality of inversion and imaging.In this paper,we solve the 3D vertical transverse isotropic(VTI)Christoffel equation and obtain the analytical,frst-order,and zero-order Taylor expansion solutions that represent P-,SV-,and SH-wave phase velocities and polarization vectors.These analytical and approximated solutions are used to build the P/S plane-wave expression identical to the far-feld term of seismic wave,from which the strain rate expressions are derived and DAS radiation patterns are thus extracted for anisotropic P/S waves.We observe that the gauge length and phase angle terms control the radiating intensity of DAS records.Additionally,the Bond transformation is adopted to derive the DAS radiation patterns in title transverse isotropic(TTI)media,which exhibits higher complexity than that of VTI media.Several synthetic examples demonstrate the feasibility and effectiveness of our theory.展开更多
The principle of optical time-domain reflection localization limits the sensing spatial resolution of Raman distributed optical fiber sensing.We provide a solution for a Raman distributed optical fiber sensing system ...The principle of optical time-domain reflection localization limits the sensing spatial resolution of Raman distributed optical fiber sensing.We provide a solution for a Raman distributed optical fiber sensing system with kilometer-level sensing distance and submeter spatial resolution.Based on this,we propose a Raman distributed optical fiber sensing scheme based on chaotic pulse cluster demodulation.Chaotic pulse clusters are used as the probe signal,in preference to conventional pulsed or chaotic single-pulse lasers.Furthermore,the accurate positioning of the temperature variety region along the sensing fiber can be realized using chaotic pulse clusters.The proposed demodulation scheme can enhance the signal-to-noise ratio by improving the correlation between the chaotic reference and the chaotic Raman anti-Stokes scattering signals.The experiment achieved a sensing spatial resolution of 30 cm at a distributed temperature-sensing distance of∼6.0 km.Furthermore,we explored the influence of chaotic pulse width and detector bandwidth on the sensing spatial resolution.In addition,the theoretical experiments proved that the sensing spatial resolution in the proposed scheme was independent of the pulse width and sensing distance.展开更多
The distribution of shear-wave velocities in the subsurface is generally used to assess the potential forseismic liquefaction and soil amplification effects and to classify seismic sites. Newly developeddistributed ac...The distribution of shear-wave velocities in the subsurface is generally used to assess the potential forseismic liquefaction and soil amplification effects and to classify seismic sites. Newly developeddistributed acoustic sensing (DAS) technology enables estimation of the shear-wave distribution as ahigh-density seismic observation system. This technology is characterized by low maintenance costs,high-resolution outputs, and real-time data transmission capabilities, albeit with the challenge ofmanaging massive data generation. Rapid and efficient interpretation of data is the key to advancingapplication of the DAS technology. In this study, field tests were carried out to record ambient noise overa short period using DAS technology, from which the surface-wave dispersion curves were extracted. Inorder to reduce the influence of directional effects on the results, an unsupervised clustering method isused to select appropriate clusters to extract the Green's function. A combination of a genetic algorithmand Monte Carlo (GA-MC) simulation is proposed to invert the subsurface velocity structure. Thestratigraphic profiles obtained by the GA-MC method are in agreement with the borehole profiles.Compared to other methods, the proposed optimization method not only improves the solution qualitybut also reduces the solution time.展开更多
Distributed fiber-optic sensing(DFOS)can turn the worldwide fiber network into a sensing array,which may immensely extend the sensing range and approaches for hazard assessment,earth observation,and human activity mea...Distributed fiber-optic sensing(DFOS)can turn the worldwide fiber network into a sensing array,which may immensely extend the sensing range and approaches for hazard assessment,earth observation,and human activity measurement.However,most existing DFOS schemes cannot simultaneously give dual attention to the detection ability(for example,sensing distance)and multipoint localizing function.A mirror-image correlation method is proposed and can precisely extract the time delay between two original signals from their composite detected signal.This method enables the distributed vibration sensing function of the laser interferometer and maintains its high detection ability.We demonstrate its feasibility by simultaneously localizing multiple knocking vibrations on a 250-km round-trip fiber and distinguishing traffic vibrations at two urban positions in a field test.The localizing precision is analyzed and satisfies the requirements for fiber network sensing.展开更多
Interferometry is a crucial investigative technique used across diverse fields to achieve highprecision measurements.It works by analyzing the phase difference between two interfering waves,which results from variatio...Interferometry is a crucial investigative technique used across diverse fields to achieve highprecision measurements.It works by analyzing the phase difference between two interfering waves,which results from variations in optical path lengths within an interferometer.We introduce a novel method for directly measuring changes in the phase difference within an optical interferometer,importantly,with the added advantage of a controllable enhancement factor.This approach is achieved through a two-step process:first,the optical phase difference is encoded into a sub-GHz radiofrequency(RF)signal using microwave-photonic manipulation;then,RF interferometry-assisted phase amplification is implemented at the destructive interference point.In our experiments,we demonstrate a phase sensitivity of 2.14 rad∕nm operating at 140 MHz using a miniature in-fiber Fabry-Pérot interferometer for sub-nanometer displacement sensing,which reveals a sensitivity magnification factor of 258.6.With further refinement,we anticipate that even higher enhancement factors can be achieved,paving the way for the development of cost-effective,ultrasensitive interferometry-based instruments for high-precision optical measurements.展开更多
Distributed acoustic sensing(DAS) is one recently developed seismic acquisition technique that is based on fiber-optic sensing. DAS provides dense spatial spacing that is useful to image shallow structure with surface...Distributed acoustic sensing(DAS) is one recently developed seismic acquisition technique that is based on fiber-optic sensing. DAS provides dense spatial spacing that is useful to image shallow structure with surface waves.To test the feasibility of DAS in shallow structure imaging,the PoroTomo team conducted a DAS experiment with the vibroseis truck T-Rex in Brady’s Hot Springs, Nevada, USA.The Rayleigh waves excited by the vertical mode of the vibroseis truck were analyzed with the Multichannel Analysis of Surface Waves(MASW) method. Phase velocities between5 and 20 Hz were successfully extracted for one segment of cable and were employed to build a shear-wave velocity model for the top 50 meters. The dispersion curves obtained with DAS agree well with the ones extracted from co-located geophones data and from the passive source Noise Correlation Functions(NCF). Comparing to the co-located geophone array, the higher sensor density that DAS arrays provides help reducing aliasing in dispersion analysis, and separating different surface wave modes. This study demonstrates the feasibility and advantage of DAS in imaging shallow structure with surface waves.展开更多
How to recognize targets with similar appearances from remote sensing images(RSIs) effectively and efficiently has become a big challenge. Recently, convolutional neural network(CNN) is preferred in the target classif...How to recognize targets with similar appearances from remote sensing images(RSIs) effectively and efficiently has become a big challenge. Recently, convolutional neural network(CNN) is preferred in the target classification due to the powerful feature representation ability and better performance. However,the training and testing of CNN mainly rely on single machine.Single machine has its natural limitation and bottleneck in processing RSIs due to limited hardware resources and huge time consuming. Besides, overfitting is a challenge for the CNN model due to the unbalance between RSIs data and the model structure.When a model is complex or the training data is relatively small,overfitting occurs and leads to a poor predictive performance. To address these problems, a distributed CNN architecture for RSIs target classification is proposed, which dramatically increases the training speed of CNN and system scalability. It improves the storage ability and processing efficiency of RSIs. Furthermore,Bayesian regularization approach is utilized in order to initialize the weights of the CNN extractor, which increases the robustness and flexibility of the CNN model. It helps prevent the overfitting and avoid the local optima caused by limited RSI training images or the inappropriate CNN structure. In addition, considering the efficiency of the Na¨?ve Bayes classifier, a distributed Na¨?ve Bayes classifier is designed to reduce the training cost. Compared with other algorithms, the proposed system and method perform the best and increase the recognition accuracy. The results show that the distributed system framework and the proposed algorithms are suitable for RSIs target classification tasks.展开更多
基金National Natural Science Foundation of China,Grant/Award Number:42130706。
文摘Coal mining induces changes in the nature of rock and soil bodies,as well as hydrogeological conditions,which can easily trigger the occurrence of geological disasters such as water inrush,movement of the coal seam roof and floor,and rock burst.Transparency in coal mine geological conditions provides technical support for intelligent coal mining and geological disaster prevention.In this sense,it is of great significance to address the requirements for informatizing coal mine geological conditions,dynamically adjust sensing parameters,and accurately identify disaster characteristics so as to prevent and control coal mine geological disasters.This paper examines the various action fields associated with geological disasters in mining faces and scrutinizes the types and sensing parameters of geological disasters resulting from coal seam mining.On this basis,it summarizes a distributed fiber-optic sensing technology framework for transparent geology in coal mines.Combined with the multi-field monitoring characteristics of the strain field,the temperature field,and the vibration field of distributed optical fiber sensing technology,parameters such as the strain increment ratio,the aquifer temperature gradient,and the acoustic wave amplitude are extracted as eigenvalues for identifying rock breaking,aquifer water level,and water cut range,and a multi-field sensing method is established for identifying the characteristics of mining-induced rock mass disasters.The development direction of transparent geology based on optical fiber sensing technology is proposed in terms of the aspects of sensing optical fiber structure for large deformation monitoring,identification accuracy of optical fiber acoustic signals,multi-parameter monitoring,and early warning methods.
基金supported by the National Key R&D Program of China(2023YFB2906303)Major Program(JD)of Hubei Province(2023BAA013)+1 种基金the National Natural Science Foundation of China(62105111,62225110)the Chilean National Agency for Research and Development(Fondecyt Regular 1241085,Fondequip EQM220113 and Basal AFB240002).
文摘Distributed fiber-optic sensing has become an indispensable tool for large-scale structural and environmentalmonitoring, where spectral interrogation of backscattering light enables high-precision quantitative measurement ofexternal perturbations. Conventional spectral analysis methods, typically based on frequency-domain serialinterrogation or time-to-frequency mapping, face inherent trade-offs between measurement speed, dynamic strainmeasurement range, and system complexity. Here, we present a distributed frequency comb enabled spectrumcorrelationreflectometry as a universal spectral analysis framework that leverages optical frequency comb for parallelmulti-frequency interrogation, which is experimentally demonstrated in a phase-sensitive optical time-domainreflectometry (φ-OTDR) system. This method eliminates the need for large frequency scans, achieving more thantenfold improvement in measurement speed over the state-of-the-art spectral analysis methods. Compared to existingphase-demodulated φ-OTDR systems, this method enables vibration amplitude monitoring with a dynamic strainmeasurement range expanded by more than an order of magnitude, while intrinsically circumventing phaseunwrapping issues and interference fading. This work establishes a new paradigm for distributed spectral analysis,providing a flexible and robust platform for a wide range of sensing technologies, including Rayleigh and Brillouinbasedschemes, which may have significant impact for geophysics, seismology, civil engineering, and other fields.
基金support of her postdoctoral research at the GFZ Helmholtz Centre for Geosciences.P.Pan acknowledges the financial support of the National Natural Science Foundation of China(Grant No.52339001)H.Hofmann and Y.Ji acknowledge the financial support of the Helmholtz Association's Initiative and Networking Fund for the Helmholtz Young Investigator Group ARES(contract number VH-NG-1516).
文摘A multi-stage stress relaxation test was performed on a granodiorite sample to understand the deformation process prior to the macroscopic failure of brittle rocks,as well as the transient response during stress relaxation.Distributed optical fiber sensing was used to measure strains across the sample surface by helically wrapping the single-mode fiber around the cylindrical sample.Close agreement was observed between the circumferential strains obtained from the optical fibers and the extensometer.The reconstructed full-field strain contours show strain heterogeneity from the crack closure phase,and the strains in the later deformation phase are dominantly localized within the former high-strain zone.The Gini coefficient was used to quantify the degree of strain localization and shows an initial increase during the crack closure phase,a decrease during the linear elastic phase,and a subsequent increase during the post-yielding phase.This behavior corresponds to a process of initial localization from an imperfect boundary condition,homogenization,and eventual relocalization prior to the macroscopic failure of the sample.The transient strain rate decay during the stress relaxation phase was quantified using the p-value in the“Omori-like"power law function.A higher initial stress at the onset of relaxation results in a lower p-value,indicating a slower strain rate decay.As the sample approaches macroscopic failure,the lowest p-value shifts from the most damaged zone to adjacent areas,suggesting stress redistribution or crack propagation in deformed crystalline rocks under stress relaxation conditions.
基金Supported by the the Academician Workstation Program of Yunnan Province(202405AF140013)High-Quality Development Special Project of the Ministry of Industry and Information Technology(TC240A9ED-56)Shanghai Agricultural Technology Innovation Project(2024-02-08-00-12-F00032)。
文摘Distributed optical fiber sensing(DOFS)technology has been widely applied in pipeline monitoring,seismic detection,and security protection due to its broad coverage,high sensitivity,and strong anti-interference capability.However,the acquired signals are typically noisy,exhibit complex temporal-spatial patterns,and contain high-dimensional categorical features,posing significant challenges for robust classification.To address these issues,this paper introduces an Inception-ResNet-based model for intrusion event recognition in DOFS systems.The Inception architecture extracts multi-scale features from complex vibration patterns,while the residual optimization of ResNet enables efficient deep feature propagation and stable training.Furthermore,to enhance model interpretability,a Grad-CAM-based mechanism is integrated to visualize class-discriminative regions in the vibration signals,revealing the patterns that most strongly influence the network's decisions.Extensive experiments demonstrate the effectiveness of the proposed approach,achieving an average classification accuracy of 92.6%,outperforming traditional deep learning networks even with significantly reduced training data.These results indicate that the interpretable Inception-ResNet framework not only accurately classifies complex one-dimensional sensing signals but also provides transparent and reliable support for practical DOFS applications.
基金supported by the National Natural Science Foundation of China(Grant No.52339001).
文摘To investigate the damage evolution caused by stress-driven and sub-critical crack propagation within the Beishan granite under multi-creep triaxial compressive conditions,the distributed optical fiber sensing and X-ray computed tomography were combined to obtain the strain distribution over the sample surface and internal fractures of the samples.The Gini and skewness(G-S)coefficients were used to quantify strain localization during tests,where the Gini coefficient reflects the degree of clustering of elements with high strain values,i.e.,strain localization/delocalization.The strain localization-induced asymmetry of data distribution is quantified by the skewness coefficient.A precursor to granite failure is defined by the rapid and simultaneous increase of the G-S coefficients,which are calculated from strain increment,giving an earlier warning of failure by about 8%peak stress than those from absolute strain values.Moreover,the process of damage accumulation due to stress-driven crack propagation in Beishan granite is different at various confining pressures as the stress exceeds the crack initiation stress.Concretely,strain localization is continuous until brittle failure at higher confining pressure,while both strain localization and delocalization occur at lower confining pressure.Despite the different stress conditions,a similar statistical characteristic of strain localization during the creep stage is observed.The Gini coefficient increases,and the skewness coefficient decreases slightly as the creep stress is below 95%peak stress.When the accelerated strain localization begins,the Gini and skewness coefficients increase rapidly and simultaneously.
基金The authors grate fully acknowledge the financial support provided by the National Natural Science Foundation of China(Grant Nos.42225702 and 42077235)the Open Research Project Program of the State Key Laboratory of Internet of Things for Smart City(University of Macao),China(Grant No.SKUoTSC(UM)-2021-2023/0RP/GA10/2022).
文摘Defects in cast-in-situ piles have an adverse impact on load transfer at the pile‒soil interface and pile bearing capacity. In recent years, thermal integrity profiling (TIP) has been developed to measure temperature profiles of cast-in-situ piles, enabling the detection of structural defects or anomalies at the early stage of construction. However, using this integrity testing method to evaluate potential defects in cast-in-situ piles requires a comprehensive understanding of the mechanism of hydration heat transfer from piles to surrounding soils. In this study, small-scale model tests were conducted in laboratory to investigate the performance of TIP in detecting pile integrity. Fiber-optic distributed temperature sensing (DTS) technology was used to monitor detailed temperature variations along model piles in sand. Additionally, sensors were installed in sand to measure water content and matric suction. An interpretation method against available DTS-based thermal profiles was proposed to reveal the potential defective regions. It shows that the temperature difference between normal and defective piles is more obvious in wet sand. In addition, there is a critical zone of water migration in sand due to the water absorption behavior of cement and temperature transfer-induced water migration in the early-age concrete setting. These findings could provide important insight into the improvement of the TIP testing method for field applications.
基金support by National Natural Science Foundation of China (Nos. 51606158, 12104402)
文摘Fiber-optic sensing technology has the advantages of passivity, anti-electromagnetic interference, longdistancemeasurement, high sensitivity and high accuracy, small size, and adaptability to harsh environments such ashigh-vacuum, high-pressure, and strong magnetic fields compared with the traditional electrical sensing technology.However, with the increasing application requirements, how to further improve the sensitivity of fiber-optic sensors,extend the detection limit and improve the maintenance-free capability has become one of the core issues of thecurrent research. This paper reviews the principle, preparation, and application of fiber-optic microstructured sensingbased on abrupt field type. It specifically outlines the development and applications of micro-nano optical fibers,photonic crystal optical fibers, optical fiber gratings and structured optical fibers, and lists the main preparationmethods of two types of micro-nano optical fibers from the basic theory of optical fiber microstructured sensordevices.
基金funding support from the National Natural Science Foundation of China(Grant No.52204030)Youth Innovation and Technology Support Program for Higher Education Institutions of Shandong Province,China(Grant No.2022KJ070)the National Natural Science Foundation of China Enterprise Innovation and Development Joint Fund Project(Grant No.U19B6003).
文摘Fiber-optic distributed strain sensing(FO-DSS)has been successful in monitoring strain changes along horizontal wellbores in hydraulically fractured reservoirs.However,the mechanism driving the various FO-DSS responses associated with near-wellbore hydraulic fracture properties is still unclear.To address this knowledge gap,we use coupled wellbore-reservoir-geomechanics simulations to study measured strain-change behavior and infer hydraulic fracture characteristics.The crossflow among fractures is captured through explicit modeling of the transient wellbore flow.In addition,local grid refinement is applied to accurately capture strain changes along the fiber.A Base Case model was designed with four fractures of varying properties,simulating strain change signals when the production well is shut-in for 10 d after 240 d of production and reopened for 2 d.Strain-pressure plots for different fracture clusters were used to gain insights into inferring fracture properties using DSS data.When comparing the model with and without the wellbore,distinct strain change signals were observed,emphasizing the importance of incorporating the wellbore in FO-DSS modeling.The effects of fracture spacing and matrix permeability on strain change signals were thoroughly investigated.The results of our numerical study can improve the understanding of the relation between DSS signals and fracture hydraulic properties,thus maximizing the value of the dataset for fracture diagnostics and characterization.
基金supported by the Science and Technology Development Plan of Jilin Province(No.20180201036GX)
文摘A distributed optical-fiber acoustic sensor is an acoustic sensor that uses the optical fiber itself as a photosensitive medium,and is based on Rayleigh backscattering in an optical fiber.The sensor is widely used in the safety monitoring of oil and gas pipelines,the classification of weak acoustic signals,defense,seismic prospecting,and other fields.In the field of seismic prospecting,distributed optical-fiber acoustic sensing(DAS)will gradually replace the use of the traditional geophone.The present paper mainly expounds the recent application of DAS,and summarizes recent research achievements of DAS in resource exploration,intrusion monitoring,pattern recognition,and other fields and various DAS system structures.It is found that the high-sensitivity and long-distance sensing capabilities of DAS play a role in the extensive monitoring applications of DAS in engineering.The future application and development of DAS technology are examined,with the hope of promoting the wider application of the DAS technology,which benefits engineering and society.
基金Supported by Key Laboratory of Space Active Optical-Electro Technology of Chinese Academy of Sciences(2021ZDKF4)。
文摘Phase-sensitive Optical Time-Domain Reflectometer(φ-OTDR)technology facilitates the real-time detection of vibration events along fiber optic cables by analyzing changes in Rayleigh scattering signals.This technology is widely used in applications such as intrusion monitoring and structural health assessments.Traditional signal processing methods,such as Support Vector Machines(SVM)and K-Nearest Neighbors(KNN),have limitations in feature extraction and classification in complex environments.Conversely,a single deep learning model often struggles with capturing long time-series dependencies and mitigating noise interference.In this study,we propose a deep learning model that integrates Convolutional Neural Network(CNN),Long Short-Term Memory Network(LSTM),and Transformer modules,leveraging φ-OTDR technology for distributed fiber vibration sensing event recognition.The hybrid model combines the CNN's capability to extract local features,the LSTM's ability to model temporal dynamics,and the Transformer's proficiency in capturing global dependencies.This integration significantly enhances the accuracy and robustness of event recognition.In experiments involving six types of vibration events,the model consistently achieved a validation accuracy of 0.92,and maintained a validation loss of approximately 0.2,surpassing other models,such as TAM+BiLSTM and CNN+CBAM.The results indicate that the CNN+LSTM+Transformer model is highly effective in handling vibration signal classification tasks in complex scenarios,offering a promising new direction for the application of fiber optic vibration sensing technology.
基金support from the Open Research Project Programme of the State Key Laboratory of Internet of Things for Smart City,University of Macao (Grant No.SKL-IoTSC (UM)-2021-2023/ORPF/A19/2022)the General Research Fund project from Research Grants Council of Hong Kong Special Administrative Region Government of China (Grant No.15214722)the Start-up Fund from The Hong Kong Polytechnic University (Grant No.BD88).
文摘Despite the extensive use of distributed fiber optic sensing(DFOS)in monitoring underground structures,its potential in detecting structural anomalies,such as cracks and cavities,is still not fully understood.To contribute to the identification of defects in underground structures,this study conducted a four-point bending test of a reinforced concrete(RC)beam and uniaxial loading tests of an RC specimen with local cavities.The experimental results revealed the disparity in DFOS strain spike profiles between these two structural anomalies.The effectiveness of DFOS in the quantification of crack opening displacement(COD)was also demonstrated,even in cases where perfect bonding was not achievable between the cable and structures.In addition,DFOS strain spikes observed in two diaphragm wall panels of a twin circular shaft were also reported.The most probable cause of those spikes was identified as the mechanical behavior associated with local concrete contamination.With the utilization of the strain profiles obtained from laboratory tests and field monitoring,three types of multi-classifiers,based on support vector machine(SVM),random forest(RF),and backpropagation neural network(BP),were employed to classify strain profiles,including crack-induced spikes,non-crack-induced spikes,and non-spike strain profiles.Among these classifiers,the SVM-based classifier exhibited superior performance in terms of accuracy and model robustness.This finding suggests that the SVM-based classifier holds promise as a potential solution for the automatic detection and classification of defects in underground structures during long-term monitoring.
基金Supported by the National High-Tech Research Project(GJSCB-HFGDY-2024-004)National Natural Science Foundation of China(12402305)+2 种基金Postdoctoral Fellowship Program of CPSF(GZC20232200)China Postdoctoral Science Foundation(2024M762703)Sichuan Science and Technology Program(2025ZNSFSC1352)。
文摘The Carter model is used to characterize the dynamic behaviors of fracture growth and fracturing fluid leakoff.A thermo-fluid coupling temperature response forward model is built considering the fluid flow and heat transfer in wellbore,fracture and reservoir.The influences of fracturing parameters and fracture parameters on the responses of distributed temperature sensing(DTS)are analyzed,and a diagnosis method of fracture parameters is presented based on the simulated annealing algorithm.A field case study is introduced to verify the model’s reliability.Typical V-shaped characteristics can be observed from the DTS responses in the multi-cluster fracturing process,with locations corresponding to the hydraulic fractures.The V-shape depth is shallower for a higher injection rate and longer fracturing and shut-in time.Also,the V-shape is wider for a higher fracture-surface leakoff coefficient,longer fracturing time and smaller fracture width.Additionally,the cooling effect near the wellbore continues to spread into the reservoir during the shut-in period,causing the DTS temperature to decrease instead of rise.Real-time monitoring and interpretation of DTS temperature data can help understand the fracture propagation during fracturing operation,so that immediate measures can be taken to improve the fracturing performance.
基金support provided by Science Foundation Ireland Frontiers for the Future Programme,21/FFP-P/10090.
文摘Curved geostructures,such as tunnels,are commonly encountered in geotechnical engineering and are critical to maintaining structural stability.Ensuring their proper performance through field monitoring during their service life is essential for the overall functionality of geotechnical infrastructure.Distributed Brillouin sensing(DBS)is increasingly applied in geotechnical projects due to its ability to acquire spatially continuous strain and temperature distributions over distances of up to 150 km using a single optical fibre.However,limited by the complex operations of distributed optic fibre sensing(DFOS)sensors in curved structures,previous reports about exploiting DBS in geotechnical structural health monitoring(SHM)have mostly been focused on flat surfaces.The lack of suitable DFOS installation methods matched to the spatial characteristics of continuous monitoring is one of the major factors that hinder the further application of this technique in curved structures.This review paper starts with a brief introduction of the fundamental working principle of DBS and the inherent limitations of DBS being used on monitoring curved surfaces.Subsequently,the state-of-the-art installation methods of optical fibres in curved structures are reviewed and compared to address the most suitable scenario of each method and their advantages and disadvantages.The installation challenges of optical fibres that can highly affect measurement accuracy are also discussed in the paper.
基金supported by the National Key R&D Program of China under grant No.2021YFA0716800。
文摘Distributed acoustic sensing(DAS)is increasingly used in seismic exploration owing to its wide frequency range,dense sampling and real-time monitoring.DAS radiation patterns help to understand angle response of DAS records and improve the quality of inversion and imaging.In this paper,we solve the 3D vertical transverse isotropic(VTI)Christoffel equation and obtain the analytical,frst-order,and zero-order Taylor expansion solutions that represent P-,SV-,and SH-wave phase velocities and polarization vectors.These analytical and approximated solutions are used to build the P/S plane-wave expression identical to the far-feld term of seismic wave,from which the strain rate expressions are derived and DAS radiation patterns are thus extracted for anisotropic P/S waves.We observe that the gauge length and phase angle terms control the radiating intensity of DAS records.Additionally,the Bond transformation is adopted to derive the DAS radiation patterns in title transverse isotropic(TTI)media,which exhibits higher complexity than that of VTI media.Several synthetic examples demonstrate the feasibility and effectiveness of our theory.
基金supported by the National Natural Science Foundation of China(Grant Nos.U23A20375 and 62075151)the National Key Research and Development Program of China(Grant No.202103021223042).
文摘The principle of optical time-domain reflection localization limits the sensing spatial resolution of Raman distributed optical fiber sensing.We provide a solution for a Raman distributed optical fiber sensing system with kilometer-level sensing distance and submeter spatial resolution.Based on this,we propose a Raman distributed optical fiber sensing scheme based on chaotic pulse cluster demodulation.Chaotic pulse clusters are used as the probe signal,in preference to conventional pulsed or chaotic single-pulse lasers.Furthermore,the accurate positioning of the temperature variety region along the sensing fiber can be realized using chaotic pulse clusters.The proposed demodulation scheme can enhance the signal-to-noise ratio by improving the correlation between the chaotic reference and the chaotic Raman anti-Stokes scattering signals.The experiment achieved a sensing spatial resolution of 30 cm at a distributed temperature-sensing distance of∼6.0 km.Furthermore,we explored the influence of chaotic pulse width and detector bandwidth on the sensing spatial resolution.In addition,the theoretical experiments proved that the sensing spatial resolution in the proposed scheme was independent of the pulse width and sensing distance.
基金supported by the National Natural Science Foundation of China(Grant Nos.42225702 and 42077235)the Natural Science Foundation of Jiangsu Province(Grant No.BK20211086)the open fund of the Key Laboratory of Earth Fissures Geological Disaster,Ministry of Natural Resources.
文摘The distribution of shear-wave velocities in the subsurface is generally used to assess the potential forseismic liquefaction and soil amplification effects and to classify seismic sites. Newly developeddistributed acoustic sensing (DAS) technology enables estimation of the shear-wave distribution as ahigh-density seismic observation system. This technology is characterized by low maintenance costs,high-resolution outputs, and real-time data transmission capabilities, albeit with the challenge ofmanaging massive data generation. Rapid and efficient interpretation of data is the key to advancingapplication of the DAS technology. In this study, field tests were carried out to record ambient noise overa short period using DAS technology, from which the surface-wave dispersion curves were extracted. Inorder to reduce the influence of directional effects on the results, an unsupervised clustering method isused to select appropriate clusters to extract the Green's function. A combination of a genetic algorithmand Monte Carlo (GA-MC) simulation is proposed to invert the subsurface velocity structure. Thestratigraphic profiles obtained by the GA-MC method are in agreement with the borehole profiles.Compared to other methods, the proposed optimization method not only improves the solution qualitybut also reduces the solution time.
基金supported by the National Natural Science Foundation of China(Grant No.62171249)the National Key R&D Program of China(Grant No.2021YFA1402102)the Tsinghua Initiative Scientific Research Program.
文摘Distributed fiber-optic sensing(DFOS)can turn the worldwide fiber network into a sensing array,which may immensely extend the sensing range and approaches for hazard assessment,earth observation,and human activity measurement.However,most existing DFOS schemes cannot simultaneously give dual attention to the detection ability(for example,sensing distance)and multipoint localizing function.A mirror-image correlation method is proposed and can precisely extract the time delay between two original signals from their composite detected signal.This method enables the distributed vibration sensing function of the laser interferometer and maintains its high detection ability.We demonstrate its feasibility by simultaneously localizing multiple knocking vibrations on a 250-km round-trip fiber and distinguishing traffic vibrations at two urban positions in a field test.The localizing precision is analyzed and satisfies the requirements for fiber network sensing.
基金support from the Roy A.Wilkens Professorship Endowment。
文摘Interferometry is a crucial investigative technique used across diverse fields to achieve highprecision measurements.It works by analyzing the phase difference between two interfering waves,which results from variations in optical path lengths within an interferometer.We introduce a novel method for directly measuring changes in the phase difference within an optical interferometer,importantly,with the added advantage of a controllable enhancement factor.This approach is achieved through a two-step process:first,the optical phase difference is encoded into a sub-GHz radiofrequency(RF)signal using microwave-photonic manipulation;then,RF interferometry-assisted phase amplification is implemented at the destructive interference point.In our experiments,we demonstrate a phase sensitivity of 2.14 rad∕nm operating at 140 MHz using a miniature in-fiber Fabry-Pérot interferometer for sub-nanometer displacement sensing,which reveals a sensitivity magnification factor of 258.6.With further refinement,we anticipate that even higher enhancement factors can be achieved,paving the way for the development of cost-effective,ultrasensitive interferometry-based instruments for high-precision optical measurements.
基金partially supported by the Geothermal Technologies Office of the USA Department of Energy (No. DE-EE0006760)the State Key Laboratory of Geodesy and Earth’s Dynamics, Institute of Geodey and Geophysics, Chinese Academy of Sciences (No. SKLGED2019-5-4-E)
文摘Distributed acoustic sensing(DAS) is one recently developed seismic acquisition technique that is based on fiber-optic sensing. DAS provides dense spatial spacing that is useful to image shallow structure with surface waves.To test the feasibility of DAS in shallow structure imaging,the PoroTomo team conducted a DAS experiment with the vibroseis truck T-Rex in Brady’s Hot Springs, Nevada, USA.The Rayleigh waves excited by the vertical mode of the vibroseis truck were analyzed with the Multichannel Analysis of Surface Waves(MASW) method. Phase velocities between5 and 20 Hz were successfully extracted for one segment of cable and were employed to build a shear-wave velocity model for the top 50 meters. The dispersion curves obtained with DAS agree well with the ones extracted from co-located geophones data and from the passive source Noise Correlation Functions(NCF). Comparing to the co-located geophone array, the higher sensor density that DAS arrays provides help reducing aliasing in dispersion analysis, and separating different surface wave modes. This study demonstrates the feasibility and advantage of DAS in imaging shallow structure with surface waves.
基金supported by the National Natural Science Foundation of China(U1435220)
文摘How to recognize targets with similar appearances from remote sensing images(RSIs) effectively and efficiently has become a big challenge. Recently, convolutional neural network(CNN) is preferred in the target classification due to the powerful feature representation ability and better performance. However,the training and testing of CNN mainly rely on single machine.Single machine has its natural limitation and bottleneck in processing RSIs due to limited hardware resources and huge time consuming. Besides, overfitting is a challenge for the CNN model due to the unbalance between RSIs data and the model structure.When a model is complex or the training data is relatively small,overfitting occurs and leads to a poor predictive performance. To address these problems, a distributed CNN architecture for RSIs target classification is proposed, which dramatically increases the training speed of CNN and system scalability. It improves the storage ability and processing efficiency of RSIs. Furthermore,Bayesian regularization approach is utilized in order to initialize the weights of the CNN extractor, which increases the robustness and flexibility of the CNN model. It helps prevent the overfitting and avoid the local optima caused by limited RSI training images or the inappropriate CNN structure. In addition, considering the efficiency of the Na¨?ve Bayes classifier, a distributed Na¨?ve Bayes classifier is designed to reduce the training cost. Compared with other algorithms, the proposed system and method perform the best and increase the recognition accuracy. The results show that the distributed system framework and the proposed algorithms are suitable for RSIs target classification tasks.
