Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm f...Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm for infrared images,F-YOLOv8,is proposed.First,a spatial-to-depth network replaces the traditional backbone network's strided convolution or pooling layer.At the same time,it combines with the channel attention mechanism so that the neural network focuses on the channels with large weight values to better extract low-resolution image feature information;then an improved feature pyramid network of lightweight bidirectional feature pyramid network(L-BiFPN)is proposed,which can efficiently fuse features of different scales.In addition,a loss function of insertion of union based on the minimum point distance(MPDIoU)is introduced for bounding box regression,which obtains faster convergence speed and more accurate regression results.Experimental results on the FLIR dataset show that the improved algorithm can accurately detect infrared road targets in real time with 3%and 2.2%enhancement in mean average precision at 50%IoU(mAP50)and mean average precision at 50%—95%IoU(mAP50-95),respectively,and 38.1%,37.3%and 16.9%reduction in the number of model parameters,the model weight,and floating-point operations per second(FLOPs),respectively.To further demonstrate the detection capability of the improved algorithm,it is tested on the public dataset PASCAL VOC,and the results show that F-YOLO has excellent generalized detection performance.展开更多
Snow depth is a general input variable in many models of agriculture,hydrology,climate and ecology.This study makes use of observational data of snow depth and explanatory variables to compare the accuracy and effect ...Snow depth is a general input variable in many models of agriculture,hydrology,climate and ecology.This study makes use of observational data of snow depth and explanatory variables to compare the accuracy and effect of geographically weighted regression kriging(GWRK)and regression kriging(RK)in a spatial interpolation of regional snow depth.The auxiliary variables are analyzed using correlation coefficients and the variance inflation factor(VIF).Three variables,Height,topographic ruggedness index(TRI),and land surface temperature(LST),are used as explanatory variables to establish a regression model for snow depth.The estimated spatial distribution of snow depth in the Bayanbulak Basin of the Tianshan Mountains in China with a spatial resolution of 1 km is obtained.The results indicate that 1)the result of GWRK's accuracy is slightly higher than that of RK(R^2=0.55 vs.R^2=0.50,RMSE(root mean square error)=0.102 m vs.RMSE=0.077 m);2)for the subareas,GWRK and RK exhibit similar estimation results of snow depth.Areas in the Bayanbulak Basin with a snow depth greater than 0.15m are mainly distributed in an elevation range of 2632.00–3269.00 m and the snow in this area comprises 45.00–46.00% of the total amount of snow in this basin.However,the GWRK resulted in more detailed information on snow depth distribution than the RK.The final conclusion is that GWRK is better suited for estimating regional snow depth distribution.展开更多
A spatial resolution effect of remote sensing bathymetry is an important scientific problem. The in situ measured water depth data and images of Dongdao Island are used to study the effect of water depth inversion fro...A spatial resolution effect of remote sensing bathymetry is an important scientific problem. The in situ measured water depth data and images of Dongdao Island are used to study the effect of water depth inversion from different spatial resolution remote sensing images. The research experiments are divided into five groups including Quick Bird and World View-2 remote sensing images with their original spatial resolution（2.4/2.0 m）and four kinds of reducing spatial resolution（4, 8, 16 and 32 m）, and the water depth control and checking points are set up to carry out remote sensing water depth inversion. The experiment results indicate that the accuracy of the water depth remote sensing inversion increases first as the spatial resolution decreases from 2.4/2.0 to 4, 8 and16 m. And then the accuracy decreases along with the decreasing spatial resolution. When the spatial resolution of the image is 16 m, the inversion error is minimum. In this case, when the spatial resolution of the remote sensing image is 16 m, the mean relative errors（MRE） of Quick Bird and World View-2 bathymetry are 21.2% and 13.1%,compared with the maximum error are decreased by 14.7% and 2.9% respectively; the mean absolute errors（MAE） are 2.0 and 1.4 m, compared with the maximum are decreased by 1.0 and 0.5 m respectively. The results provide an important reference for the selection of remote sensing data in the study and application of the remote sensing bathymetry.展开更多
This paper presents a novel approach to model and simulate the multi-support depth-varying seismic motions（MDSMs） within heterogeneous offshore and onshore sites.Based on 1 D wave propagation theory,the three-dimens...This paper presents a novel approach to model and simulate the multi-support depth-varying seismic motions（MDSMs） within heterogeneous offshore and onshore sites.Based on 1 D wave propagation theory,the three-dimensional ground motion transfer functions on the surface or within an offshore or onshore site are derived by considering the effects of seawater and porous soils on the propagation of seismic P waves.Moreover,the depth-varying and spatial variation properties of seismic ground motions are considered in the ground motion simulation.Using the obtained transfer functions at any locations within a site,the offshore or onshore depth-varying seismic motions are stochastically simulated based on the spectral representation method（SRM）.The traditional approaches for simulating spatially varying ground motions are improved and extended to generate MDSMs within multiple offshore and onshore sites.The simulation results show that the PSD functions and coherency losses of the generated MDSMs are compatible with respective target values,which fully validates the effectiveness of the proposed simulation method.The synthesized MDSMs can provide strong support for the precise seismic response prediction and performance-based design of both offshore and onshore large-span engineering structures.展开更多
This article adopts three artificial intelligence techniques, Gaussian Process Regression（GPR）, Least Square Support Vector Machine（LSSVM） and Extreme Learning Machine（ELM）, for prediction of rock depth（d） at ...This article adopts three artificial intelligence techniques, Gaussian Process Regression（GPR）, Least Square Support Vector Machine（LSSVM） and Extreme Learning Machine（ELM）, for prediction of rock depth（d） at any point in Chennai. GPR, ELM and LSSVM have been used as regression techniques.Latitude and longitude are also adopted as inputs of the GPR, ELM and LSSVM models. The performance of the ELM, GPR and LSSVM models has been compared. The developed ELM, GPR and LSSVM models produce spatial variability of rock depth and offer robust models for the prediction of rock depth.展开更多
The aquatic eco-environment is significantly affected by temporal and spatial variation of the mixed layer depth （MLD） in large shallow lakes. In the present study, we simulated the three-dimensional water temperatu...The aquatic eco-environment is significantly affected by temporal and spatial variation of the mixed layer depth （MLD） in large shallow lakes. In the present study, we simulated the three-dimensional water temperature of Taihu Lake with an unstructured grid with a finite-volume coastal ocean model （FVCOM） using wind speed, wind direction, short-wave radiation and other meteorological data measured during 13-18 August 2008. The simulated results were consistent with the measurements. The temporal and spatial distribution of the MLD and the possible relevant mechanisms were analyzed on the basis of the water temperature profile data of Taihu Lake. The results indicated that diurnal stratification might be established through the combined effect of the hydrodynamic conditions induced by wind and the heat exchange between air and water. Compared with the net heat flux, the changes of the MLD were delayed approximately two hours. Furthermore, there were significant spatial differences of the MLD in Taihu Lake due to the combined impact of thermal and hydrodynamic forces. Briefly, diurnal stratification formed relatively easily in Gonghu Bay, Zhushan Bay, Xukou Bay and East Taihu Bay, and the surface mixed layer was thin. The center of the lake region had the deepest surface mixed layer due to the strong mixing process. In addition, Meiliang Bay showed a medium depth of the surface mixed layer. Our analysis indicated that the spatial difference in the hydrodynamic action was probably the major cause for the spatial variation of the MLD in Taihu Lake.展开更多
Recently developed time-of-flight principle based depth-sensing video camera technologies provide precise per-pixel range data in addition to color video. Such cameras will find application in robotics and vision-base...Recently developed time-of-flight principle based depth-sensing video camera technologies provide precise per-pixel range data in addition to color video. Such cameras will find application in robotics and vision-based human computer interaction scenarios such as games and gesture input systems. Time-of-flight principle range cameras are becoming more and more available. They promise to make the 3D reconstruction of scenes easier, avoiding the practical issues resulting from 3D imaging techniques based on triangulation or disparity estimation. A spatial touch system was presented which uses a depth-sensing camera to touch spatial objects and details on its implementation, and how this technology will enable new spatial interactions was speculated.展开更多
By using the observational snow data of more than 700 weather stations,the interannual temporal and spatial characteristics of seasonal snow cover in China were analyzed.The results show that northern Xinjiang,northea...By using the observational snow data of more than 700 weather stations,the interannual temporal and spatial characteristics of seasonal snow cover in China were analyzed.The results show that northern Xinjiang,northeastern China–Inner Mongolia,and the southwestern and southern portions of Tibetan Plateau are three regions in China with high seasonal snow cover and also an interannual anomaly of snow cover.According to the trend of both the snow depth and snow cover days,there are three changing patterns for the seasonal snow cover:The first type is that both snow depth and snow cover days simultaneously increase or decrease;this includes northern Xinjiang,middle and eastern Inner Mongolia,and so on.The second is that snow depth increases but snow cover days decrease;this type mainly locates in the eastern parts of the northeastern plain of China and the upper reaches of the Yangtze River.The last type is that snow depth decreases but snow cover days increase at the same time such as that in middle parts of Tibetan Plateau.Snow cover in China appears to have been having a slow increasing trend during the last 40 years.On the decadal scale,snow depth and snow cover days slightly increased in the 1960s and then decreased in the 1970s;they again turn to increasing in the 1980s and persist into 1990s.展开更多
基金supported by the National Natural Science Foundation of China(No.62103298)。
文摘Aiming at the problems of low detection accuracy and large model size of existing object detection algorithms applied to complex road scenes,an improved you only look once version 8(YOLOv8)object detection algorithm for infrared images,F-YOLOv8,is proposed.First,a spatial-to-depth network replaces the traditional backbone network's strided convolution or pooling layer.At the same time,it combines with the channel attention mechanism so that the neural network focuses on the channels with large weight values to better extract low-resolution image feature information;then an improved feature pyramid network of lightweight bidirectional feature pyramid network(L-BiFPN)is proposed,which can efficiently fuse features of different scales.In addition,a loss function of insertion of union based on the minimum point distance(MPDIoU)is introduced for bounding box regression,which obtains faster convergence speed and more accurate regression results.Experimental results on the FLIR dataset show that the improved algorithm can accurately detect infrared road targets in real time with 3%and 2.2%enhancement in mean average precision at 50%IoU(mAP50)and mean average precision at 50%—95%IoU(mAP50-95),respectively,and 38.1%,37.3%and 16.9%reduction in the number of model parameters,the model weight,and floating-point operations per second(FLOPs),respectively.To further demonstrate the detection capability of the improved algorithm,it is tested on the public dataset PASCAL VOC,and the results show that F-YOLO has excellent generalized detection performance.
基金supported by Projects of International Cooperation and Exchanges NSFC (grant: 41361140361)the Special fund project of Chinese Academy of Sciences (grant: Y371164001)the key deployment project of Chinese Academy of Sciences (Grant No. KZZD-EW-12-2, KZZD-EW12-3)
文摘Snow depth is a general input variable in many models of agriculture,hydrology,climate and ecology.This study makes use of observational data of snow depth and explanatory variables to compare the accuracy and effect of geographically weighted regression kriging(GWRK)and regression kriging(RK)in a spatial interpolation of regional snow depth.The auxiliary variables are analyzed using correlation coefficients and the variance inflation factor(VIF).Three variables,Height,topographic ruggedness index(TRI),and land surface temperature(LST),are used as explanatory variables to establish a regression model for snow depth.The estimated spatial distribution of snow depth in the Bayanbulak Basin of the Tianshan Mountains in China with a spatial resolution of 1 km is obtained.The results indicate that 1)the result of GWRK's accuracy is slightly higher than that of RK(R^2=0.55 vs.R^2=0.50,RMSE(root mean square error)=0.102 m vs.RMSE=0.077 m);2)for the subareas,GWRK and RK exhibit similar estimation results of snow depth.Areas in the Bayanbulak Basin with a snow depth greater than 0.15m are mainly distributed in an elevation range of 2632.00–3269.00 m and the snow in this area comprises 45.00–46.00% of the total amount of snow in this basin.However,the GWRK resulted in more detailed information on snow depth distribution than the RK.The final conclusion is that GWRK is better suited for estimating regional snow depth distribution.
基金The National Key Technology Research and Development Program of China under contract No.2012BAB16B01
文摘A spatial resolution effect of remote sensing bathymetry is an important scientific problem. The in situ measured water depth data and images of Dongdao Island are used to study the effect of water depth inversion from different spatial resolution remote sensing images. The research experiments are divided into five groups including Quick Bird and World View-2 remote sensing images with their original spatial resolution（2.4/2.0 m）and four kinds of reducing spatial resolution（4, 8, 16 and 32 m）, and the water depth control and checking points are set up to carry out remote sensing water depth inversion. The experiment results indicate that the accuracy of the water depth remote sensing inversion increases first as the spatial resolution decreases from 2.4/2.0 to 4, 8 and16 m. And then the accuracy decreases along with the decreasing spatial resolution. When the spatial resolution of the image is 16 m, the inversion error is minimum. In this case, when the spatial resolution of the remote sensing image is 16 m, the mean relative errors（MRE） of Quick Bird and World View-2 bathymetry are 21.2% and 13.1%,compared with the maximum error are decreased by 14.7% and 2.9% respectively; the mean absolute errors（MAE） are 2.0 and 1.4 m, compared with the maximum are decreased by 1.0 and 0.5 m respectively. The results provide an important reference for the selection of remote sensing data in the study and application of the remote sensing bathymetry.
基金National Key R&D Program of China under Grant No.2016YFC0701108the State Key Program of National Natural Science Foundation of China under Grant No.51738007
文摘This paper presents a novel approach to model and simulate the multi-support depth-varying seismic motions（MDSMs） within heterogeneous offshore and onshore sites.Based on 1 D wave propagation theory,the three-dimensional ground motion transfer functions on the surface or within an offshore or onshore site are derived by considering the effects of seawater and porous soils on the propagation of seismic P waves.Moreover,the depth-varying and spatial variation properties of seismic ground motions are considered in the ground motion simulation.Using the obtained transfer functions at any locations within a site,the offshore or onshore depth-varying seismic motions are stochastically simulated based on the spectral representation method（SRM）.The traditional approaches for simulating spatially varying ground motions are improved and extended to generate MDSMs within multiple offshore and onshore sites.The simulation results show that the PSD functions and coherency losses of the generated MDSMs are compatible with respective target values,which fully validates the effectiveness of the proposed simulation method.The synthesized MDSMs can provide strong support for the precise seismic response prediction and performance-based design of both offshore and onshore large-span engineering structures.
文摘This article adopts three artificial intelligence techniques, Gaussian Process Regression（GPR）, Least Square Support Vector Machine（LSSVM） and Extreme Learning Machine（ELM）, for prediction of rock depth（d） at any point in Chennai. GPR, ELM and LSSVM have been used as regression techniques.Latitude and longitude are also adopted as inputs of the GPR, ELM and LSSVM models. The performance of the ELM, GPR and LSSVM models has been compared. The developed ELM, GPR and LSSVM models produce spatial variability of rock depth and offer robust models for the prediction of rock depth.
基金supported by the National Natoral Science Foundation of Ghina (Grant Nos.41071070.41165001.and 41171368)the Special Scientific Research Project of the China Commonwealth Trade(meteorology)(GYHY201006054)
文摘The aquatic eco-environment is significantly affected by temporal and spatial variation of the mixed layer depth （MLD） in large shallow lakes. In the present study, we simulated the three-dimensional water temperature of Taihu Lake with an unstructured grid with a finite-volume coastal ocean model （FVCOM） using wind speed, wind direction, short-wave radiation and other meteorological data measured during 13-18 August 2008. The simulated results were consistent with the measurements. The temporal and spatial distribution of the MLD and the possible relevant mechanisms were analyzed on the basis of the water temperature profile data of Taihu Lake. The results indicated that diurnal stratification might be established through the combined effect of the hydrodynamic conditions induced by wind and the heat exchange between air and water. Compared with the net heat flux, the changes of the MLD were delayed approximately two hours. Furthermore, there were significant spatial differences of the MLD in Taihu Lake due to the combined impact of thermal and hydrodynamic forces. Briefly, diurnal stratification formed relatively easily in Gonghu Bay, Zhushan Bay, Xukou Bay and East Taihu Bay, and the surface mixed layer was thin. The center of the lake region had the deepest surface mixed layer due to the strong mixing process. In addition, Meiliang Bay showed a medium depth of the surface mixed layer. Our analysis indicated that the spatial difference in the hydrodynamic action was probably the major cause for the spatial variation of the MLD in Taihu Lake.
基金supported by Ministry of Knowl-edge Economy(MKE), Korea as a project,"The nextgeneration core technology for Intelligent Informationand electronics"
文摘Recently developed time-of-flight principle based depth-sensing video camera technologies provide precise per-pixel range data in addition to color video. Such cameras will find application in robotics and vision-based human computer interaction scenarios such as games and gesture input systems. Time-of-flight principle range cameras are becoming more and more available. They promise to make the 3D reconstruction of scenes easier, avoiding the practical issues resulting from 3D imaging techniques based on triangulation or disparity estimation. A spatial touch system was presented which uses a depth-sensing camera to touch spatial objects and details on its implementation, and how this technology will enable new spatial interactions was speculated.
文摘By using the observational snow data of more than 700 weather stations,the interannual temporal and spatial characteristics of seasonal snow cover in China were analyzed.The results show that northern Xinjiang,northeastern China–Inner Mongolia,and the southwestern and southern portions of Tibetan Plateau are three regions in China with high seasonal snow cover and also an interannual anomaly of snow cover.According to the trend of both the snow depth and snow cover days,there are three changing patterns for the seasonal snow cover:The first type is that both snow depth and snow cover days simultaneously increase or decrease;this includes northern Xinjiang,middle and eastern Inner Mongolia,and so on.The second is that snow depth increases but snow cover days decrease;this type mainly locates in the eastern parts of the northeastern plain of China and the upper reaches of the Yangtze River.The last type is that snow depth decreases but snow cover days increase at the same time such as that in middle parts of Tibetan Plateau.Snow cover in China appears to have been having a slow increasing trend during the last 40 years.On the decadal scale,snow depth and snow cover days slightly increased in the 1960s and then decreased in the 1970s;they again turn to increasing in the 1980s and persist into 1990s.
