In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To addr...In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To address this issue,we integrate Large Kernel Convolution(LKconv)into the U-Net framework,proposing an enhanced network structure named U-LKconv network,which significantly enhances the capability to recover image details even under low sampling conditions.展开更多
Conventional shot-gather migration uses a cross-correlation imaging condition proposed by Clarebout （1971）, which cannot preserve imaging amplitudes. The deconvolution imaging condition can improve the imaging ampli...Conventional shot-gather migration uses a cross-correlation imaging condition proposed by Clarebout （1971）, which cannot preserve imaging amplitudes. The deconvolution imaging condition can improve the imaging amplitude and compensate for illumination. However, the deconvolution imaging condition introduces instability issues. The least-squares imaging condition first computes the sum of the cross-correlation of the forward and backward wavefields over all frequencies and sources, and then divides the result by the total energy of the forward wavefield. Therefore, the least-squares imaging condition is more stable than the classic imaging condition. However, the least-squares imaging condition cannot provide accurate results in areas where the illumination is very poor and unbalanced. To stabilize the least-squares imaging condition and balance the imaging amplitude, we propose a novel imaging condition with structure constraints that is based on the least-squares imaging condition. Our novel imaging condition uses a plane wave construction that constrains the imaging result to be smooth along geological structure boundaries in the inversion frame. The proposed imaging condition improves the stability of the imaging condition and balances the imaging amplitude. The proposed condition is applied to two examples, the horizontal layered model and the Sigsbee 2A model. These tests show that, in comparison to the damped least-squares imaging condition, the stabilized least-squares imaging condition with structure constraints improves illumination stability and balance, makes events more consecutive, adjusts the amplitude of the depth layers where the illumination is poor and unbalanced, suppresses imaging artifacts, and is conducive to amplitude preserving imaging of deep layers.展开更多
Clustering earthquakes refer to the seismic events that occur closely with each other in time and space. Because their overlapping waveform records make it difficult to pick the first arrivals, the hypocenters of clus...Clustering earthquakes refer to the seismic events that occur closely with each other in time and space. Because their overlapping waveform records make it difficult to pick the first arrivals, the hypocenters of clustering earthquakes cannot be determined accurately by traveltime location methods. Here we apply a reverse-time imaging （RTI） method to map clustering earthquakes. Taking the advantage of directly using waveforms, the RTI method is capable to map either a single small earthquake or some densely distributed clustering earthquakes beneath a 2-D seismic array. In 3-D case the RTI method is successfully applied to locate the long-offset doublet earthquakes using the data from a set of sparsely distributed surface stations. However, for the same acquisition geometry, the RTI encounters challenges in mapping densely distributed clustering earthquakes. While it is obvious that improving the mapping of clustering earthquakes requires a denser receiver network with wider range of illumination angles, it is necessary to verify the actual resolution of the RTI method with synthetic data. In our study area in the Three Gorges region, our tests in 3-D case suggest that some events beneath the linear aligned sub-arrays have reasonable resolution.展开更多
With the increasing complexity of prospecting objectives,reverse time migration( RTM) has attracted more and more attention due to its outstanding imaging quality. RTM is based on two-way wave equation,so it can avoid...With the increasing complexity of prospecting objectives,reverse time migration( RTM) has attracted more and more attention due to its outstanding imaging quality. RTM is based on two-way wave equation,so it can avoid the limits of angle in traditional one-way wave equation migration,image reverse branch,prism waves and multi-reflected wave precisely and obtain accurate dynamic information. However,the huge demands for storage and computation as well as low frequency noises restrict its wide application. The normalized cross-correlation imaging conditions based on wave field decomposition are derived from traditional cross-correlation imaging condition,and it can eliminate the low-frequency noises effectively and improve the imaging resolution. The practical procedure includes separating source and receiver wave field into one-way components respectively,and conducting cross-correlation imaging condition to the post-separated wave field. In this way,the resolution and precision of the imaging result will be promoted greatly.展开更多
Elastic reverse-time migration can effectively deal with multicomponent seismic data in which the imaging condition based on energy norm can extract the scalar-imaging result from multicomponent data.However,the energ...Elastic reverse-time migration can effectively deal with multicomponent seismic data in which the imaging condition based on energy norm can extract the scalar-imaging result from multicomponent data.However,the energy cross-correlation imaging condition characterized by particle velocity and stress suffers from the problem of overdependence on the background elastic parameters.Therefore,we characterize the elastic-wave energy using the energy-flow vector,which is equal to the energy density,without background elastic parameters.According to the source and receiver wave fields,we propose an imaging energyflow vector and an elastic-wave energy imaging condition.Under the assumption of a planewave solution,the backscattering suppression is verified.The numerical simulations show that the elastic-energy imaging condition can obtain the energy image without backscattering.Compared with the cross-correlation imaging conditions in a vector-based wave field,the proposed imaging condition can eliminate the dependence on the background elastic parameters and effectively process seabed multicomponent data,which are conducive to further providing an interpretation of marine geological structures.展开更多
The reverse time migration(RTM)of ground penetrating radar(GPR)is usually implemented in its two-dimensional(2D)form,due to huge computational cost.However,2D RTM algorithm is difficult to focus the scattering signal ...The reverse time migration(RTM)of ground penetrating radar(GPR)is usually implemented in its two-dimensional(2D)form,due to huge computational cost.However,2D RTM algorithm is difficult to focus the scattering signal and produce a high precision subsurface image when the object is buried in a complicated subsurface environment.To better handle the multi-off set GPR data,we propose a three-dimensional(3D)prestack RTM algorithm.The high-order fi nite diff erence time domian(FDTD)method,with the accuracy of eighth-order in space and second-order in time,is applied to simulate the forward and backward extrapolation electromagnetic fi elds.In addition,we use the normalized correlation imaging condition to obtain pre-stack RTM result and the Laplace fi lter to suppress the low frequency noise generated during the correlation process.The numerical test of 3D simulated GPR data demonstrated that 3D RTM image shows excellent coincidence with the true model.Compared with 2D RTM image,the 3D RTM image can more clearly and accurately refl ect the 3D spatial distribution of the target,and the resolution of the imaging results is far better.Furthermore,the application of observed GPR data further validates the eff ectiveness of the proposed 3D GPR RTM algorithm,and its fi nal image can more reliably guide the subsequent interpretation.展开更多
Background:Visual conditioning can refine the response of neurons in the visual cortex and higher visual and cognitive processing of a presented stimulus.This process results in increased sensitivity for that stimulus...Background:Visual conditioning can refine the response of neurons in the visual cortex and higher visual and cognitive processing of a presented stimulus.This process results in increased sensitivity for that stimulus.The development of new optical imaging technology in the field of neuroscience has led to important advances,notably to better define the functional organization and plasticity of visual areas.The objective of this project is to determine the effect of daily visual conditioning with an oblique sinusoidal grating on the distribution and amplitude of cortical responses.For this,we use wide-field calcium imaging on awake mice,allowing for the observation of responses to a stimulus throughout the entire cortex in real time.Methods:C57BL/6 mice,expressing the GCaMP6s calcium reporter gene,are used to longitudinally measure neuronal activity via wide-field calcium imaging.Spontaneous activity at rest,as well as cortical responses to visual stimuli consisting of sinusoidal networks with orientation(0,30°,60°and 90°),spatial frequency(0.03,0.12,0.24 and 0.48 cpd)and contrast(100%,75%and 50%)variables are recorded to establish cortical maps,as well as tuning curves.Subsequently,the baseline function of the cortex,as well as the cortical representation of visual stimulation(30°or 90°,0.03 cpd and a contrast of 50%,75%and 100%)are studied in the animal before,during,and after daily monocular conditioning,consisting of a specific sinusoidal network(30°,0.03 cpd and 100%)over a period of 7 days.The variations in intensity and activation specificity of various visual cortical areas are calculated according to the visual conditioning and compared to an orientation stimulus for which the animal has not been conditioned(90°).Results:The cortical activation curves show a greater sensitivity of response for stimuli having horizontal or vertical gratings(0 and 90°)than for oblique gratings(30°and 60°)at low spatial frequencies(0,0.3 and 0.12 cpd).However,this trend does not occur with high spatial frequencies(0.24 and 0.48 cpd).Finally,although the intensity of activation varies in a way that is not proportional to the contrast of the stimulation,it would have no influence on the perception of the orientation of the stimuli.Conditioning at a 30°stimulus results in greater activation of the primary visual cortex and some extra-striate visual areas,as well as greater amplification of the ipsilateral cortical responses to the presentation of the visual stimuli.Conclusions:In conclusion,the results demonstrate that visual conditioning would allow for plasticity and consolidation of higher visual pathways.展开更多
This paper presents a new method for refining image annotation by integrating probabilistic la- tent semantic analysis （PLSA） with conditional random field （CRF）. First a PLSA model with asymmetric modalities is c...This paper presents a new method for refining image annotation by integrating probabilistic la- tent semantic analysis （PLSA） with conditional random field （CRF）. First a PLSA model with asymmetric modalities is constructed to predict a candidate set of annotations with confidence scores, and then model semantic relationship among the candidate annotations by leveraging conditional ran- dom field. In CRF, the confidence scores generated lay the PLSA model and the Fliekr distance be- tween pairwise candidate annotations are considered as local evidences and contextual potentials re- spectively. The novelty of our method mainly lies in two aspects ： exploiting PLSA to predict a candi- date set of annotations with confidence scores as well as CRF to further explore the semantic context among candidate annotations for precise image annotation. To demonstrate the effectiveness of the method proposed in this paper, an experiment is conducted on the standard Corel dataset and its re- sults are ＇compared favorably with several state-of-the-art approaches.展开更多
To reduce the computation cost of a combined probabilistic graphical model and a deep neural network in semantic segmentation, the local region condition random field (LRCRF) model is investigated which selectively ap...To reduce the computation cost of a combined probabilistic graphical model and a deep neural network in semantic segmentation, the local region condition random field (LRCRF) model is investigated which selectively applies the condition random field (CRF) to the most active region in the image. The full convolutional network structure is optimized with the ResNet-18 structure and dilated convolution to expand the receptive field. The tracking networks are also improved based on SiameseFC by considering the frame relations in consecutive-frame traffic scene maps. Moreover, the segmentation results of the greyscale input data sets are more stable and effective than using the RGB images for deep neural network feature extraction. The experimental results show that the proposed method takes advantage of the image features directly and achieves good real-time performance and high segmentation accuracy.展开更多
Currently,deep convolutional neural networks have made great progress in the field of semantic segmentation.Because of the fixed convolution kernel geometry,standard convolution neural networks have been limited the a...Currently,deep convolutional neural networks have made great progress in the field of semantic segmentation.Because of the fixed convolution kernel geometry,standard convolution neural networks have been limited the ability to simulate geometric transformations.Therefore,a deformable convolution is introduced to enhance the adaptability of convolutional networks to spatial transformation.Considering that the deep convolutional neural networks cannot adequately segment the local objects at the output layer due to using the pooling layers in neural network architecture.To overcome this shortcoming,the rough prediction segmentation results of the neural network output layer will be processed by fully connected conditional random fields to improve the ability of image segmentation.The proposed method can easily be trained by end-to-end using standard backpropagation algorithms.Finally,the proposed method is tested on the ISPRS dataset.The results show that the proposed method can effectively overcome the influence of the complex structure of the segmentation object and obtain state-of-the-art accuracy on the ISPRS Vaihingen 2D semantic labeling dataset.展开更多
Wear particles are inevitably occurred from moving parts, such as a piston-cylinder made from steel or hybrid materials. And a durability of these parts must be evaluated. The wear particle analysis has been known as ...Wear particles are inevitably occurred from moving parts, such as a piston-cylinder made from steel or hybrid materials. And a durability of these parts must be evaluated. The wear particle analysis has been known as a very effective method to foreknow and decide a moving situation and a damage of machine parts by using the digital computer image processing. But it is not laid down to calculate shape parameters of wear particle and wear volume. In order to apply image processing method in a durability evaluation of machine parts, it needs to verify the reliability of the calculated data by the image processing and to lay down the number of images and the amount of wear particles in one image. In this work, the lubricated friction experiment was carried out in order to establish the optimum image capture with the 1045 specimen under experiment condition. The wear particle data were calculated differently according to the number of image and the amount of wear particle in one image. The results show that capturing conditions need to be more than 140 wear particles in one image and over 40 images for the reliable data. Thus, the capturing method of wear particles images was optimized for condition diagnosis of machine moving parts.展开更多
Cork oak in Maamora forest is experiencing the dieback phenomenon. The evaluation of the latter in this forest has gained the importance over time and with the solicitation of managers to objectify its phytosanitary s...Cork oak in Maamora forest is experiencing the dieback phenomenon. The evaluation of the latter in this forest has gained the importance over time and with the solicitation of managers to objectify its phytosanitary situation. Aiming at prioritizing management actions, remote sensing seems to be an effective tool to inquire about stands’ health conditions and their evolution. To this end, this study aims at mapping and validating health status of cork oak stands in Maamora. Sentinel 2 images in 2015 and 2020 were processed to calculate the differential normalized difference water index (NDWI), revealing vegetation moisture variation caused by drought. A statistical method based on thresholds was used to map cork oak dieback stands, those with no changes and those recovered. Results have shown that 54.63% of cork oak in Maamora forest have not changed in terms of phytosanitary situation between 2015 and 2020, 31.10% of oak stands are afflicted by a slight decline and 12.97% by a severe decline. Areas with slight or strong recovery remain minimal and represent 1.04% and 0.25% respectively. Ground data indicated that the map generated displayed a good distinction between stands severely and slightly declined with a global accuracy of 66.66%. Therefore, further research elaborating an advanced vegetation index reflecting the various factors of dieback would be of much importance.展开更多
The research of removing rain from pictures or videos has always been an important topic in the field of computer vision and image processing. Most noise reduction methods more or less remove texture details in rain-f...The research of removing rain from pictures or videos has always been an important topic in the field of computer vision and image processing. Most noise reduction methods more or less remove texture details in rain-free areas, resulting in an over-smoothing effect in the restored background. The research on image noise removal is very meaningful. We exploit the powerful generative power of a modified generative adversarial network (CGAN) by enforcing an additional condition that makes the derained image indistinguishable from its corresponding ground-truth clean image. An efficient and lightweight attention machine mechanism NAM is introduced in the generator, and an IDN-CGAN model is proposed to capture image salient features through attention operations. Taking advantage of the mutual information in different dimensions of the features to further suppress insignificant channels or pixels to ensure better visual quality, we also introduce a new fine-grained loss function in the generator-discriminator pair, predicting and real data degree of disparity to achieve improved results.展开更多
Based on arbitrarily wide-angle wave equations,a reverse-time propagation scheme is developed by substituting the partial derivatives of depth and time with central differences. The partial derivative of horizontal di...Based on arbitrarily wide-angle wave equations,a reverse-time propagation scheme is developed by substituting the partial derivatives of depth and time with central differences. The partial derivative of horizontal direction is replaced with high order difference. The imaging condition is computed by solving the eikonal equations. On the basis of above techniques,a prestack reverse-time depth migration algorithm is developed. The processing exam-ples of synthetic data show that the method can remove unwanted internal reflections and decrease the migration noise. The method also has the advantage of fidelity and is applicable of dip angle reflector imaging.展开更多
The imaging of offset VSP data in local phase space can improve the image of the subsurface structure near the well.In this paper,we present a migration scheme for imaging VSP data in a local phase space,which uses th...The imaging of offset VSP data in local phase space can improve the image of the subsurface structure near the well.In this paper,we present a migration scheme for imaging VSP data in a local phase space,which uses the Gabor-Daubechies tight framebased extrapolator（G-D extrapolator） and its high-frequency asymptotic expansion to extrapolate wavefields and also delineates an improved correlation imaging condition in the local angle domain.The results for migrating synthetic and real VSP data demonstrate that the application of the high-frequency G-D extrapolator asymptotic expansion can effectively decrease computational complexity.The local angle domain correlation imaging condition can be used to weaken migration artifacts without increasing computation.展开更多
AIM:To investigate the predictive value of narrowband imaging with magnifying endoscopy (NBI-ME) for identifying gastric intestinal metaplasia (GIM) in unselected patients. METHODS:We prospectively evaluated consecuti...AIM:To investigate the predictive value of narrowband imaging with magnifying endoscopy (NBI-ME) for identifying gastric intestinal metaplasia (GIM) in unselected patients. METHODS:We prospectively evaluated consecutive patients undergoing upper endoscopy for various indications, such as epigastric discomfort/pain, anaemia, gastro-oesophageal reflux disease, suspicion of peptic ulcer disease, or chronic liver diseases. Patients underwent NBI-ME, which was performed by three blinded, experienced endoscopists. In addition, five biopsies (2 antrum, 1 angulus, and 2 corpus) were taken and examined by two pathologists unaware of the endoscopic findings to determine the presence or absence of GIM. The correlation between light blue crest (LBC) appearance and histology was measured. Moreover, we quantified the degree of LBC appearance as less than 20% (+), 20%-80% (++) and more than 80% (+++) of an image field, and the semiquantitative evaluation of LBC appearance was correlated with IM percentage from the histological findings. RESULTS:We enrolled 100 (58 F/42 M) patients who were mainly referred for gastro-esophageal reflux disease/dyspepsia (46%), cancer screening/anaemia (34%), chronic liver disease (9%), and suspected celiac disease (6%); the remaining patients were referred for other indications. The prevalence of Helicobacter pylori (H. pylori ) infection detected from the biopsies was 31%, while 67% of the patients used proton pump inhibitors. LBCs were found in the antrum of 33 patients (33%); 20 of the cases were classified as LBC+, 9 as LBC++, and 4 as LBC+++. LBCs were found in the gastric body of 6 patients (6%), with 5 of them also having LBCs in the antrum. The correlation between the appearance of LBCs and histological GIM was good, with a sensitivity of 80% (95%CI:67-92), a specificity of 96% (95%CI:93-99), a positive predictive value of 84% (95%CI:73-96), a negative predictive value of 95% (95%CI:92-98), and an accuracy of 93% (95%CI:90-97). The NBI-ME examination overlooked GIM in 8 cases, but the GIM was less than 5% in 7 of the cases. Moreover, in the 6 false positive cases, the histological examination showed the presence of reactive gastropathy (4 cases) or H. pylori active chronic gastritis (2 cases). The semiquantitative correlation between the rate of LBC appearance and the percentage of GIM was 79% (P < 0.01). CONCLUSION:NBI-ME achieved good sensitivity and specificity in recognising GIM in an unselected population. In routine clinical practice, this technique can reliably target gastric biopsies.展开更多
Correctly locating the tunnel lining cavity is extremely important tunnel quality inspection.High-accuracy imaging results are hard to obtain because conventional one-way wave migration is greatly aff ected by lateral...Correctly locating the tunnel lining cavity is extremely important tunnel quality inspection.High-accuracy imaging results are hard to obtain because conventional one-way wave migration is greatly aff ected by lateral velocity change and inclination limitation and because the diff racted wave cannot be accurately returned to the real spatial position of the lining cavity.This paper presents a tunnel lining cavity imaging method based on the groundpenetrating radar(GPR)reverse-time migration(RTM)algorithm.The principle of GPR RTM is described in detail using the electromagnetic wave equation.The finite-difference timedomain method is employed to calculate the backward extrapolation electromagnetic fi elds,and the zero-time imaging condition based on the exploding-reflector concept is used to obtain the RTM results.On this basis,the GPR RTM program is compiled and applied to the simulated and observed GPR data of a typical tunnel lining cavity GPR model and a physical lining cavity model.Comparison of RTM and Kirchhoff migration results reveals that the RTM can better converge the diff racted waves of steel bar and cavity to their true position and have higher resolution and better suppress the eff ect of multiple interference and clutter scattering waves.In addition,comparison of RTM results of diff erent degrees of noise shows that RTM has strong anti-interference ability and can be used for the accurate interpretation of radar profi le in a strong interference environment.展开更多
Imaging the PP- and PS-wave for the elastic vector wave reverse-time migration requires separating the P- and S-waves during the wave field extrapolation. The amplitude and phase of the P- and S-waves are distorted wh...Imaging the PP- and PS-wave for the elastic vector wave reverse-time migration requires separating the P- and S-waves during the wave field extrapolation. The amplitude and phase of the P- and S-waves are distorted when divergence and curl operators are used to separate the P- and S-waves. We present a P- and S-wave amplitude-preserving separation algorithm for the elastic wavefield extrapolation. First, we add the P-wave pressure and P-wave vibration velocity equation to the conventional elastic wave equation to decompose the P- and S-wave vectors. Then, we synthesize the scalar P- and S-wave from the vector P- and S-wave to obtain the scalar P- and S-wave. The amplitude-preserved separated P- and S-waves are imaged based on the vector wave reverse-time migration (RTM). This method ensures that the amplitude and phase of the separated P- and S-wave remain unchanged compared with the divergence and curl operators. In addition, after decomposition, the P-wave pressure and vibration velocity can be used to suppress the interlayer reflection noise and to correct the S-wave polarity. This improves the image quality of P- and S-wave in multicomponent seismic data and the true-amplitude elastic reverse time migration used in prestack inversion.展开更多
Diffusion tensor MRI （DT-MRI or DTI） is emerging as an important non-invasive technology for elucidating intemal brain structures. It has recently been utilized to diagnose a series of diseases that affect the integ...Diffusion tensor MRI （DT-MRI or DTI） is emerging as an important non-invasive technology for elucidating intemal brain structures. It has recently been utilized to diagnose a series of diseases that affect the integrity of neural systems to provide a basis for neuroregenerative studies. Results from the present study suggested that neural tissue is reconstructed with multiple diffusion-weighted gradient directions DTI, which varies from traditional imaging methods that utilize 6 gradient directions. Simultaneously, the diffusion tensor matrix is obtained by multiple linear regressions from an equation of echo signal intensity. The condition number value and standard deviation of fractional anisotropy for each scheme can be used to evaluate image quality. Results demonstrated that increasing gradient direction to some extent resulted in improved effects. Therefore, the traditional 6 and 15 directions should not be considered optimal scan protocols for clinical DTI application. In a scheme with 20 directions, the condition number and standard deviation of fractional anisotropy of the encoding gradients matrix were significantly reduced, and resulted in more clearly and accurately displayed neural tissue. Results demonstrated that the scheme with 20 diffusion gradient directions provided better accuracy of structural renderings and could be an optimal scan protocol for clinical DTI application.展开更多
文摘In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To address this issue,we integrate Large Kernel Convolution(LKconv)into the U-Net framework,proposing an enhanced network structure named U-LKconv network,which significantly enhances the capability to recover image details even under low sampling conditions.
基金financially supported by Important National Science and Technology Specific Projects of China(Grant No. 2011ZX05023-005-005)
文摘Conventional shot-gather migration uses a cross-correlation imaging condition proposed by Clarebout （1971）, which cannot preserve imaging amplitudes. The deconvolution imaging condition can improve the imaging amplitude and compensate for illumination. However, the deconvolution imaging condition introduces instability issues. The least-squares imaging condition first computes the sum of the cross-correlation of the forward and backward wavefields over all frequencies and sources, and then divides the result by the total energy of the forward wavefield. Therefore, the least-squares imaging condition is more stable than the classic imaging condition. However, the least-squares imaging condition cannot provide accurate results in areas where the illumination is very poor and unbalanced. To stabilize the least-squares imaging condition and balance the imaging amplitude, we propose a novel imaging condition with structure constraints that is based on the least-squares imaging condition. Our novel imaging condition uses a plane wave construction that constrains the imaging result to be smooth along geological structure boundaries in the inversion frame. The proposed imaging condition improves the stability of the imaging condition and balances the imaging amplitude. The proposed condition is applied to two examples, the horizontal layered model and the Sigsbee 2A model. These tests show that, in comparison to the damped least-squares imaging condition, the stabilized least-squares imaging condition with structure constraints improves illumination stability and balance, makes events more consecutive, adjusts the amplitude of the depth layers where the illumination is poor and unbalanced, suppresses imaging artifacts, and is conducive to amplitude preserving imaging of deep layers.
基金supported by the National Natural Science Foundation of China (Nos.41230318,41204087,and 41304109)the Natural Science Foundation of Shandong Province (No.ZR2014DM006)+1 种基金the Specialized Research Fund for the Doctoral Program of Higher Education of China (No.20130132110023) the Key Laboratory of Marine Hydrocarbon Resources and Environmental Geology,Ministry of Land and Resources of China (No.MRE201303)
文摘Clustering earthquakes refer to the seismic events that occur closely with each other in time and space. Because their overlapping waveform records make it difficult to pick the first arrivals, the hypocenters of clustering earthquakes cannot be determined accurately by traveltime location methods. Here we apply a reverse-time imaging （RTI） method to map clustering earthquakes. Taking the advantage of directly using waveforms, the RTI method is capable to map either a single small earthquake or some densely distributed clustering earthquakes beneath a 2-D seismic array. In 3-D case the RTI method is successfully applied to locate the long-offset doublet earthquakes using the data from a set of sparsely distributed surface stations. However, for the same acquisition geometry, the RTI encounters challenges in mapping densely distributed clustering earthquakes. While it is obvious that improving the mapping of clustering earthquakes requires a denser receiver network with wider range of illumination angles, it is necessary to verify the actual resolution of the RTI method with synthetic data. In our study area in the Three Gorges region, our tests in 3-D case suggest that some events beneath the linear aligned sub-arrays have reasonable resolution.
文摘With the increasing complexity of prospecting objectives,reverse time migration( RTM) has attracted more and more attention due to its outstanding imaging quality. RTM is based on two-way wave equation,so it can avoid the limits of angle in traditional one-way wave equation migration,image reverse branch,prism waves and multi-reflected wave precisely and obtain accurate dynamic information. However,the huge demands for storage and computation as well as low frequency noises restrict its wide application. The normalized cross-correlation imaging conditions based on wave field decomposition are derived from traditional cross-correlation imaging condition,and it can eliminate the low-frequency noises effectively and improve the imaging resolution. The practical procedure includes separating source and receiver wave field into one-way components respectively,and conducting cross-correlation imaging condition to the post-separated wave field. In this way,the resolution and precision of the imaging result will be promoted greatly.
基金supported by the National Nature Science Foundation of China(No.61801275)Shangdong Provincial Natural Science Foundation(No.ZR2018BF002)+2 种基金China Postdoctoral Science Foundation(No.2017M622242)Basic Research Projects of Science,Education and Industry Integration Pilot Project of Qilu University of Technology(2022PX082)Qingdao Applied Research Projects.
文摘Elastic reverse-time migration can effectively deal with multicomponent seismic data in which the imaging condition based on energy norm can extract the scalar-imaging result from multicomponent data.However,the energy cross-correlation imaging condition characterized by particle velocity and stress suffers from the problem of overdependence on the background elastic parameters.Therefore,we characterize the elastic-wave energy using the energy-flow vector,which is equal to the energy density,without background elastic parameters.According to the source and receiver wave fields,we propose an imaging energyflow vector and an elastic-wave energy imaging condition.Under the assumption of a planewave solution,the backscattering suppression is verified.The numerical simulations show that the elastic-energy imaging condition can obtain the energy image without backscattering.Compared with the cross-correlation imaging conditions in a vector-based wave field,the proposed imaging condition can eliminate the dependence on the background elastic parameters and effectively process seabed multicomponent data,which are conducive to further providing an interpretation of marine geological structures.
基金This work is supported by the National Natural Science Foundation of China(No.41604039,41604102,41764005,41574078)Guangxi Natural Science Foundation project(No.2020GXNSFAA159121,2016GXNSFBA380215).
文摘The reverse time migration(RTM)of ground penetrating radar(GPR)is usually implemented in its two-dimensional(2D)form,due to huge computational cost.However,2D RTM algorithm is difficult to focus the scattering signal and produce a high precision subsurface image when the object is buried in a complicated subsurface environment.To better handle the multi-off set GPR data,we propose a three-dimensional(3D)prestack RTM algorithm.The high-order fi nite diff erence time domian(FDTD)method,with the accuracy of eighth-order in space and second-order in time,is applied to simulate the forward and backward extrapolation electromagnetic fi elds.In addition,we use the normalized correlation imaging condition to obtain pre-stack RTM result and the Laplace fi lter to suppress the low frequency noise generated during the correlation process.The numerical test of 3D simulated GPR data demonstrated that 3D RTM image shows excellent coincidence with the true model.Compared with 2D RTM image,the 3D RTM image can more clearly and accurately refl ect the 3D spatial distribution of the target,and the resolution of the imaging results is far better.Furthermore,the application of observed GPR data further validates the eff ectiveness of the proposed 3D GPR RTM algorithm,and its fi nal image can more reliably guide the subsequent interpretation.
文摘Background:Visual conditioning can refine the response of neurons in the visual cortex and higher visual and cognitive processing of a presented stimulus.This process results in increased sensitivity for that stimulus.The development of new optical imaging technology in the field of neuroscience has led to important advances,notably to better define the functional organization and plasticity of visual areas.The objective of this project is to determine the effect of daily visual conditioning with an oblique sinusoidal grating on the distribution and amplitude of cortical responses.For this,we use wide-field calcium imaging on awake mice,allowing for the observation of responses to a stimulus throughout the entire cortex in real time.Methods:C57BL/6 mice,expressing the GCaMP6s calcium reporter gene,are used to longitudinally measure neuronal activity via wide-field calcium imaging.Spontaneous activity at rest,as well as cortical responses to visual stimuli consisting of sinusoidal networks with orientation(0,30°,60°and 90°),spatial frequency(0.03,0.12,0.24 and 0.48 cpd)and contrast(100%,75%and 50%)variables are recorded to establish cortical maps,as well as tuning curves.Subsequently,the baseline function of the cortex,as well as the cortical representation of visual stimulation(30°or 90°,0.03 cpd and a contrast of 50%,75%and 100%)are studied in the animal before,during,and after daily monocular conditioning,consisting of a specific sinusoidal network(30°,0.03 cpd and 100%)over a period of 7 days.The variations in intensity and activation specificity of various visual cortical areas are calculated according to the visual conditioning and compared to an orientation stimulus for which the animal has not been conditioned(90°).Results:The cortical activation curves show a greater sensitivity of response for stimuli having horizontal or vertical gratings(0 and 90°)than for oblique gratings(30°and 60°)at low spatial frequencies(0,0.3 and 0.12 cpd).However,this trend does not occur with high spatial frequencies(0.24 and 0.48 cpd).Finally,although the intensity of activation varies in a way that is not proportional to the contrast of the stimulation,it would have no influence on the perception of the orientation of the stimuli.Conditioning at a 30°stimulus results in greater activation of the primary visual cortex and some extra-striate visual areas,as well as greater amplification of the ipsilateral cortical responses to the presentation of the visual stimuli.Conclusions:In conclusion,the results demonstrate that visual conditioning would allow for plasticity and consolidation of higher visual pathways.
基金Supported by the National Basic Research Priorities Programme(No.2013CB329502)the National High Technology Research and Development Programme of China(No.2012AA011003)+1 种基金the Natural Science Basic Research Plan in Shanxi Province of China(No.2014JQ2-6036)the Science and Technology R&D Program of Baoji City(No.203020013,2013R2-2)
文摘This paper presents a new method for refining image annotation by integrating probabilistic la- tent semantic analysis （PLSA） with conditional random field （CRF）. First a PLSA model with asymmetric modalities is constructed to predict a candidate set of annotations with confidence scores, and then model semantic relationship among the candidate annotations by leveraging conditional ran- dom field. In CRF, the confidence scores generated lay the PLSA model and the Fliekr distance be- tween pairwise candidate annotations are considered as local evidences and contextual potentials re- spectively. The novelty of our method mainly lies in two aspects ： exploiting PLSA to predict a candi- date set of annotations with confidence scores as well as CRF to further explore the semantic context among candidate annotations for precise image annotation. To demonstrate the effectiveness of the method proposed in this paper, an experiment is conducted on the standard Corel dataset and its re- sults are ＇compared favorably with several state-of-the-art approaches.
文摘To reduce the computation cost of a combined probabilistic graphical model and a deep neural network in semantic segmentation, the local region condition random field (LRCRF) model is investigated which selectively applies the condition random field (CRF) to the most active region in the image. The full convolutional network structure is optimized with the ResNet-18 structure and dilated convolution to expand the receptive field. The tracking networks are also improved based on SiameseFC by considering the frame relations in consecutive-frame traffic scene maps. Moreover, the segmentation results of the greyscale input data sets are more stable and effective than using the RGB images for deep neural network feature extraction. The experimental results show that the proposed method takes advantage of the image features directly and achieves good real-time performance and high segmentation accuracy.
基金National Key Research and Development Program of China(No.2017YFC0405806)。
文摘Currently,deep convolutional neural networks have made great progress in the field of semantic segmentation.Because of the fixed convolution kernel geometry,standard convolution neural networks have been limited the ability to simulate geometric transformations.Therefore,a deformable convolution is introduced to enhance the adaptability of convolutional networks to spatial transformation.Considering that the deep convolutional neural networks cannot adequately segment the local objects at the output layer due to using the pooling layers in neural network architecture.To overcome this shortcoming,the rough prediction segmentation results of the neural network output layer will be processed by fully connected conditional random fields to improve the ability of image segmentation.The proposed method can easily be trained by end-to-end using standard backpropagation algorithms.Finally,the proposed method is tested on the ISPRS dataset.The results show that the proposed method can effectively overcome the influence of the complex structure of the segmentation object and obtain state-of-the-art accuracy on the ISPRS Vaihingen 2D semantic labeling dataset.
基金Project supported by Research Funds from Dong-A University,Korea
文摘Wear particles are inevitably occurred from moving parts, such as a piston-cylinder made from steel or hybrid materials. And a durability of these parts must be evaluated. The wear particle analysis has been known as a very effective method to foreknow and decide a moving situation and a damage of machine parts by using the digital computer image processing. But it is not laid down to calculate shape parameters of wear particle and wear volume. In order to apply image processing method in a durability evaluation of machine parts, it needs to verify the reliability of the calculated data by the image processing and to lay down the number of images and the amount of wear particles in one image. In this work, the lubricated friction experiment was carried out in order to establish the optimum image capture with the 1045 specimen under experiment condition. The wear particle data were calculated differently according to the number of image and the amount of wear particle in one image. The results show that capturing conditions need to be more than 140 wear particles in one image and over 40 images for the reliable data. Thus, the capturing method of wear particles images was optimized for condition diagnosis of machine moving parts.
文摘Cork oak in Maamora forest is experiencing the dieback phenomenon. The evaluation of the latter in this forest has gained the importance over time and with the solicitation of managers to objectify its phytosanitary situation. Aiming at prioritizing management actions, remote sensing seems to be an effective tool to inquire about stands’ health conditions and their evolution. To this end, this study aims at mapping and validating health status of cork oak stands in Maamora. Sentinel 2 images in 2015 and 2020 were processed to calculate the differential normalized difference water index (NDWI), revealing vegetation moisture variation caused by drought. A statistical method based on thresholds was used to map cork oak dieback stands, those with no changes and those recovered. Results have shown that 54.63% of cork oak in Maamora forest have not changed in terms of phytosanitary situation between 2015 and 2020, 31.10% of oak stands are afflicted by a slight decline and 12.97% by a severe decline. Areas with slight or strong recovery remain minimal and represent 1.04% and 0.25% respectively. Ground data indicated that the map generated displayed a good distinction between stands severely and slightly declined with a global accuracy of 66.66%. Therefore, further research elaborating an advanced vegetation index reflecting the various factors of dieback would be of much importance.
文摘The research of removing rain from pictures or videos has always been an important topic in the field of computer vision and image processing. Most noise reduction methods more or less remove texture details in rain-free areas, resulting in an over-smoothing effect in the restored background. The research on image noise removal is very meaningful. We exploit the powerful generative power of a modified generative adversarial network (CGAN) by enforcing an additional condition that makes the derained image indistinguishable from its corresponding ground-truth clean image. An efficient and lightweight attention machine mechanism NAM is introduced in the generator, and an IDN-CGAN model is proposed to capture image salient features through attention operations. Taking advantage of the mutual information in different dimensions of the features to further suppress insignificant channels or pixels to ensure better visual quality, we also introduce a new fine-grained loss function in the generator-discriminator pair, predicting and real data degree of disparity to achieve improved results.
文摘Based on arbitrarily wide-angle wave equations,a reverse-time propagation scheme is developed by substituting the partial derivatives of depth and time with central differences. The partial derivative of horizontal direction is replaced with high order difference. The imaging condition is computed by solving the eikonal equations. On the basis of above techniques,a prestack reverse-time depth migration algorithm is developed. The processing exam-ples of synthetic data show that the method can remove unwanted internal reflections and decrease the migration noise. The method also has the advantage of fidelity and is applicable of dip angle reflector imaging.
基金supported by the National Hi-Tech Research and Development Program of China (Grant No.2006AA09A102-11)the National Natural Science Fund of China (Grant No.40730424 and 40674064)
文摘The imaging of offset VSP data in local phase space can improve the image of the subsurface structure near the well.In this paper,we present a migration scheme for imaging VSP data in a local phase space,which uses the Gabor-Daubechies tight framebased extrapolator（G-D extrapolator） and its high-frequency asymptotic expansion to extrapolate wavefields and also delineates an improved correlation imaging condition in the local angle domain.The results for migrating synthetic and real VSP data demonstrate that the application of the high-frequency G-D extrapolator asymptotic expansion can effectively decrease computational complexity.The local angle domain correlation imaging condition can be used to weaken migration artifacts without increasing computation.
文摘AIM:To investigate the predictive value of narrowband imaging with magnifying endoscopy (NBI-ME) for identifying gastric intestinal metaplasia (GIM) in unselected patients. METHODS:We prospectively evaluated consecutive patients undergoing upper endoscopy for various indications, such as epigastric discomfort/pain, anaemia, gastro-oesophageal reflux disease, suspicion of peptic ulcer disease, or chronic liver diseases. Patients underwent NBI-ME, which was performed by three blinded, experienced endoscopists. In addition, five biopsies (2 antrum, 1 angulus, and 2 corpus) were taken and examined by two pathologists unaware of the endoscopic findings to determine the presence or absence of GIM. The correlation between light blue crest (LBC) appearance and histology was measured. Moreover, we quantified the degree of LBC appearance as less than 20% (+), 20%-80% (++) and more than 80% (+++) of an image field, and the semiquantitative evaluation of LBC appearance was correlated with IM percentage from the histological findings. RESULTS:We enrolled 100 (58 F/42 M) patients who were mainly referred for gastro-esophageal reflux disease/dyspepsia (46%), cancer screening/anaemia (34%), chronic liver disease (9%), and suspected celiac disease (6%); the remaining patients were referred for other indications. The prevalence of Helicobacter pylori (H. pylori ) infection detected from the biopsies was 31%, while 67% of the patients used proton pump inhibitors. LBCs were found in the antrum of 33 patients (33%); 20 of the cases were classified as LBC+, 9 as LBC++, and 4 as LBC+++. LBCs were found in the gastric body of 6 patients (6%), with 5 of them also having LBCs in the antrum. The correlation between the appearance of LBCs and histological GIM was good, with a sensitivity of 80% (95%CI:67-92), a specificity of 96% (95%CI:93-99), a positive predictive value of 84% (95%CI:73-96), a negative predictive value of 95% (95%CI:92-98), and an accuracy of 93% (95%CI:90-97). The NBI-ME examination overlooked GIM in 8 cases, but the GIM was less than 5% in 7 of the cases. Moreover, in the 6 false positive cases, the histological examination showed the presence of reactive gastropathy (4 cases) or H. pylori active chronic gastritis (2 cases). The semiquantitative correlation between the rate of LBC appearance and the percentage of GIM was 79% (P < 0.01). CONCLUSION:NBI-ME achieved good sensitivity and specificity in recognising GIM in an unselected population. In routine clinical practice, this technique can reliably target gastric biopsies.
基金supported by the National Natural Science Foundation of China (Nos. 41764005, 41604039, 41604102, and 41574078)Guangxi Natural Science Foundation of China (Nos. 2016GXNSFBA380082 and 2016GXNSFBA380215)+2 种基金Guangxi Young and Middle-aged Teacher Basic Ability Improvement Project (No. KY2016YB199)Guangxi Collaborative Innovation Center for Exploration of Hidden Nonferrous Metal Deposits and Development of New Materials Project (No. GXYSXTZX2017-II-5)Guangxi Scholarship Fund of Guangxi Education Department。
文摘Correctly locating the tunnel lining cavity is extremely important tunnel quality inspection.High-accuracy imaging results are hard to obtain because conventional one-way wave migration is greatly aff ected by lateral velocity change and inclination limitation and because the diff racted wave cannot be accurately returned to the real spatial position of the lining cavity.This paper presents a tunnel lining cavity imaging method based on the groundpenetrating radar(GPR)reverse-time migration(RTM)algorithm.The principle of GPR RTM is described in detail using the electromagnetic wave equation.The finite-difference timedomain method is employed to calculate the backward extrapolation electromagnetic fi elds,and the zero-time imaging condition based on the exploding-reflector concept is used to obtain the RTM results.On this basis,the GPR RTM program is compiled and applied to the simulated and observed GPR data of a typical tunnel lining cavity GPR model and a physical lining cavity model.Comparison of RTM and Kirchhoff migration results reveals that the RTM can better converge the diff racted waves of steel bar and cavity to their true position and have higher resolution and better suppress the eff ect of multiple interference and clutter scattering waves.In addition,comparison of RTM results of diff erent degrees of noise shows that RTM has strong anti-interference ability and can be used for the accurate interpretation of radar profi le in a strong interference environment.
基金supported by Special Research Grant for Non-profit Public Service(No.201511037)National Natural Science Foundation of China(No.41504109,41506084,and 41406071)+1 种基金China Postdoctoral Science Foundation(No.2015M582060)Qingdao Municipal Applied Research Projects(No.2015308)
文摘Imaging the PP- and PS-wave for the elastic vector wave reverse-time migration requires separating the P- and S-waves during the wave field extrapolation. The amplitude and phase of the P- and S-waves are distorted when divergence and curl operators are used to separate the P- and S-waves. We present a P- and S-wave amplitude-preserving separation algorithm for the elastic wavefield extrapolation. First, we add the P-wave pressure and P-wave vibration velocity equation to the conventional elastic wave equation to decompose the P- and S-wave vectors. Then, we synthesize the scalar P- and S-wave from the vector P- and S-wave to obtain the scalar P- and S-wave. The amplitude-preserved separated P- and S-waves are imaged based on the vector wave reverse-time migration (RTM). This method ensures that the amplitude and phase of the separated P- and S-wave remain unchanged compared with the divergence and curl operators. In addition, after decomposition, the P-wave pressure and vibration velocity can be used to suppress the interlayer reflection noise and to correct the S-wave polarity. This improves the image quality of P- and S-wave in multicomponent seismic data and the true-amplitude elastic reverse time migration used in prestack inversion.
基金supported by the National Natural Science Foundation of China (Key technology of neural fiber reconstruction based on MRI),No. 60703045
文摘Diffusion tensor MRI （DT-MRI or DTI） is emerging as an important non-invasive technology for elucidating intemal brain structures. It has recently been utilized to diagnose a series of diseases that affect the integrity of neural systems to provide a basis for neuroregenerative studies. Results from the present study suggested that neural tissue is reconstructed with multiple diffusion-weighted gradient directions DTI, which varies from traditional imaging methods that utilize 6 gradient directions. Simultaneously, the diffusion tensor matrix is obtained by multiple linear regressions from an equation of echo signal intensity. The condition number value and standard deviation of fractional anisotropy for each scheme can be used to evaluate image quality. Results demonstrated that increasing gradient direction to some extent resulted in improved effects. Therefore, the traditional 6 and 15 directions should not be considered optimal scan protocols for clinical DTI application. In a scheme with 20 directions, the condition number and standard deviation of fractional anisotropy of the encoding gradients matrix were significantly reduced, and resulted in more clearly and accurately displayed neural tissue. Results demonstrated that the scheme with 20 diffusion gradient directions provided better accuracy of structural renderings and could be an optimal scan protocol for clinical DTI application.
