Estimating depth from images captured by camera sensors is crucial for the advancement of autonomous driving technologies and has gained significant attention in recent years.However,most previous methods rely on stac...Estimating depth from images captured by camera sensors is crucial for the advancement of autonomous driving technologies and has gained significant attention in recent years.However,most previous methods rely on stacked pooling or stride convolution to extract high-level features,which can limit network performance and lead to information redundancy.This paper proposes an improved bidirectional feature pyramid module(BiFPN)and a channel attention module(Seblock:squeeze and excitation)to address these issues in existing methods based on monocular camera sensor.The Seblock redistributes channel feature weights to enhance useful information,while the improved BiFPN facilitates efficient fusion of multi-scale features.The proposed method is in an end-to-end solution without any additional post-processing,resulting in efficient depth estimation.Experiment results show that the proposed method is competitive with state-of-the-art algorithms and preserves fine-grained texture of scene depth.展开更多
Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input t...Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input to augment the RGB images.Depth-based methods attempt to convert estimated depth maps to pseudo-LiDAR and then use LiDAR-based object detectors or focus on the perspective of image and depth fusion learning.However,they demonstrate limited performance and efficiency as a result of depth inaccuracy and complex fusion mode with convolutions.Different from these approaches,our proposed depth-guided vision transformer with a normalizing flows(NF-DVT)network uses normalizing flows to build priors in depth maps to achieve more accurate depth information.Then we develop a novel Swin-Transformer-based backbone with a fusion module to process RGB image patches and depth map patches with two separate branches and fuse them using cross-attention to exchange information with each other.Furthermore,with the help of pixel-wise relative depth values in depth maps,we develop new relative position embeddings in the cross-attention mechanism to capture more accurate sequence ordering of input tokens.Our method is the first Swin-Transformer-based backbone architecture for monocular 3D object detection.The experimental results on the KITTI and the challenging Waymo Open datasets show the effectiveness of our proposed method and superior performance over previous counterparts.展开更多
Joint calibration of sensors is an important prerequisite in intelligent driving scene retrieval and recognition. A simple and efficient solution is proposed for solving the problem of automatic joint calibration regi...Joint calibration of sensors is an important prerequisite in intelligent driving scene retrieval and recognition. A simple and efficient solution is proposed for solving the problem of automatic joint calibration registration between the monocular camera and the 16-line lidar. The study is divided into two parts: single-sensor independent calibration and multi-sensor joint registration, in which the selected objective world is used. The system associates the lidar coordinates with the camera coordinates. The lidar and the camera are used to obtain the normal vectors of the calibration plate and the point cloud data representing the calibration plate by the appropriate algorithm. Iterated closest points(ICP) is the method used for the iterative refinement of the registration.展开更多
The rotation matrix estimation problem is a keypoint for mobile robot localization, navigation, and control. Based on the quaternion theory and the epipolar geometry, an extended Kalman filter (EKF) algorithm is propo...The rotation matrix estimation problem is a keypoint for mobile robot localization, navigation, and control. Based on the quaternion theory and the epipolar geometry, an extended Kalman filter (EKF) algorithm is proposed to estimate the rotation matrix by using a single-axis gyroscope and the image points correspondence from a monocular camera. The experimental results show that the precision of mobile robot s yaw angle estimated by the proposed EKF algorithm is much better than the results given by the image-only and gyroscope-only method, which demonstrates that our method is a preferable way to estimate the rotation for the autonomous mobile robot applications.展开更多
A novel and fast three-dimensional reconstruction method for a Compton camera and its performance in radionuclide imaging is proposed and analyzed in this study. The conical surface sampling back-projection method wit...A novel and fast three-dimensional reconstruction method for a Compton camera and its performance in radionuclide imaging is proposed and analyzed in this study. The conical surface sampling back-projection method with scattering angle correction(CSS-BP-SC) can quickly perform the back-projection process of the Compton cone and can be used to precompute the list-mode maximum likelihood expectation maximization(LM-MLEM). A dedicated parallel architecture was designed for the graphics processing unit acceleration of the back-projection and iteration stage of the CSS-BP-SC-based LM-MLEM. The imaging results of the two-point source Monte Carlo(MC) simulation demonstrate that by analyzing the full width at half maximum along the three coordinate axes, the CSS-BP-SC-based LM-MLEM can obtain imaging results comparable to those of the traditional reconstruction algorithm, that is, the simple back-projection-based LM-MLEM. The imaging results of the mouse phantom MC simulation and experiment demonstrate that the reconstruction results obtained by the proposed method sufficiently coincide with the set radioactivity distribution, and the speed increased by more than 664 times compared to the traditional reconstruction algorithm in the mouse phantom experiment. The proposed method will further advance the imaging applications of Compton cameras.展开更多
Compton camera-based prompt gamma(PG) imaging has been proposed for range verification during proton therapy. However, a deviation between the PG and dose distributions, as well as the difference between the reconstru...Compton camera-based prompt gamma(PG) imaging has been proposed for range verification during proton therapy. However, a deviation between the PG and dose distributions, as well as the difference between the reconstructed PG and exact values, limit the effectiveness of the approach in accurate range monitoring during clinical applications. The aim of the study was to realize a PG-based dose reconstruction with a Compton camera, thereby further improving the prediction accuracy of in vivo range verification and providing a novel method for beam monitoring during proton therapy. In this paper, we present an approach based on a subset-driven origin ensemble with resolution recovery and a double evolutionary algorithm to reconstruct the dose depth profile(DDP) from the gamma events obtained by a cadmium-zinc-telluride Compton camera with limited position and energy resolution. Simulations of proton pencil beams with clinical particle rate irradiating phantoms made of different materials and the CT-based thoracic phantom were used to evaluate the feasibility of the proposed method. The results show that for the monoenergetic proton pencil beam irradiating homogeneous-material box phantom,the accuracy of the reconstructed DDP was within 0.3 mm for range prediction and within 5.2% for dose prediction. In particular, for 1.6-Gy irradiation in the therapy simulation of thoracic tumors, the range deviation of the reconstructed spreadout Bragg peak was within 0.8 mm, and the relative dose deviation in the peak area was less than 7% compared to the exact values. The results demonstrate the potential and feasibility of the proposed method in future Compton-based accurate dose reconstruction and range verification during proton therapy.展开更多
This paper proposes a person-following method based on monocular vision,which allows quadruped robots to track a target person in both indoor and outdoor environments with different illumination conditions.Our method ...This paper proposes a person-following method based on monocular vision,which allows quadruped robots to track a target person in both indoor and outdoor environments with different illumination conditions.Our method is composed of a person detector,a Kalman filter(KF)tracker,and a re-identification module.To be more specific,the person detector uses a human pose estimation method to detect persons.The KF is applied to predict the position of the target person and update its state with detection results.A re-identification module is proposed to deal with distractions,where the Convolutional Channel Features(CCF)is used to extract appearance features and Online Boosting is used to distinguish the target person from others.Especially,we design a target recapture mechanism based on the Recurrent Neural Network(RNN).Combining motion information with appearance features,the system can accurately re-identify the target person.Without extra customized markers,our method can track the target person steadily in real-time only using a monocular camera.Experiments results can validate the robustness and effectiveness of the proposed method.展开更多
Lanthanum bromide(LaBr_(3))crystal has a high energy resolution and time resolution and has been used in Compton cameras(CCs)over the past few decades.However,LaBr_(3) crystal arrays are difficult to process because L...Lanthanum bromide(LaBr_(3))crystal has a high energy resolution and time resolution and has been used in Compton cameras(CCs)over the past few decades.However,LaBr_(3) crystal arrays are difficult to process because LaBr_(3) is easy to crack and break;thus,few LaBr_(3)-based CC prototypes have been built.In this study,we designed and fabricated a large-pixel LaBr_(3) CC prototype and evaluated its performance with regard to position,energy,and angular resolution.We used two 10×10 LaBr_(3) crystal arrays with a pixel size of 5 mm×5 mm,silicon photomultipliers(SiPMs),and corresponding decoding circuits to construct our prototype.Additionally,a framework based on a Voronoi diagram and a lookup table was developed for list-mode projection data acquisition.Monte Carlo(MC)simulations based on Geant4 and experiments were conducted to evaluate the performance of our CC prototype.The lateral position resolution was 5 mm,and the maximum deviation in the depth direction was 2.5 and 5 mm for the scatterer and absorber,respectively.The corresponding measured energy resolu-tions were 7.65%and 8.44%,respectively,at 511 keV.The experimental results of ^(137)Cs point-like sources were consistent with the MC simulation results with regard to the spatial positions and full widths at half maximum(FWHMs).The angular resolution of the fabricated prototype was approximately 6°when a point-like ^(137)Cs source was centrally placed at a distance of 5 cm from the scatterer.We proposed and investigated a large-pixel LaBr_(3) CC for the first time and verified its feasibility for use in accurate spatial positioning of radiative sources with a high angular resolution.The proposed CC can satisfy the requirements of radiative source imaging and positioning in the nuclear industry and medical applications.展开更多
Background Based on the seminal work proposed by Zhou et al., much of the recent progress in learning monocular visual odometry, i.e., depth and camera motion from monocular videos, can be attributed to the tricks in ...Background Based on the seminal work proposed by Zhou et al., much of the recent progress in learning monocular visual odometry, i.e., depth and camera motion from monocular videos, can be attributed to the tricks in the training procedure, such as data augmentation and learning objectives. Methods Herein, we categorize a collection of such tricks through the theoretical examination and empirical evaluation of their effects on the final accuracy of the visual odometry. Results/Conclusions By combining the aforementioned tricks, we were able to significantly improve a baseline model adapted from SfMLearner without additional inference costs. Furthermore, we analyzed the principles of these tricks and the reason for their success. Practical guidelines for future research are also presented.展开更多
Theγ-rays are widely and abundantly present in strong nuclear radiation environments,and when they act on the camera equipment used to obtain environmental visual information on nuclear robots,radiation effects will ...Theγ-rays are widely and abundantly present in strong nuclear radiation environments,and when they act on the camera equipment used to obtain environmental visual information on nuclear robots,radiation effects will occur,which will degrade the performance of the camera system,reduce the imaging quality,and even cause catastrophic consequences.Color reducibility is an important index for evaluating the imaging quality of color camera,but its degradation mechanism in a nuclear radiation environment is still unclear.In this paper,theγ-ray irradiation experiments of CMOS cameras were carried out to analyse the degradation law of the camera’s color reducibility with cumulative irradiation and reveal the degradation mechanism of the color information of the CMOS camera underγ-ray irradiation.The results show that the spectral response of CMOS image sensor(CIS)and the spectral transmittance of lens after irradiation affect the values of a^(*)and b^(*)in the LAB color model.While the full well capacity(FWC)of CIS and transmittance of lens affect the value of L^(*)in the LAB color model,thus increase color difference and reduce brightness,the combined effect of color difference and brightness degradation will reduce the color reducibility of CMOS cameras.Therefore,the degradation of the color information of the CMOS camera afterγ-ray irradiation mainly comes from the changes in the FWC and spectral response of CIS,and the spectral transmittance of lens.展开更多
The high bandwidth and low latency of 6G network technology enable the successful application of monocular 3D object detection on vehicle platforms.Monocular 3D-object-detection-based Pseudo-LiDAR is a low-cost,lowpow...The high bandwidth and low latency of 6G network technology enable the successful application of monocular 3D object detection on vehicle platforms.Monocular 3D-object-detection-based Pseudo-LiDAR is a low-cost,lowpower solution compared to LiDAR solutions in the field of autonomous driving.However,this technique has some problems,i.e.,(1)the poor quality of generated Pseudo-LiDAR point clouds resulting from the nonlinear error distribution of monocular depth estimation and(2)the weak representation capability of point cloud features due to the neglected global geometric structure features of point clouds existing in LiDAR-based 3D detection networks.Therefore,we proposed a Pseudo-LiDAR confidence sampling strategy and a hierarchical geometric feature extraction module for monocular 3D object detection.We first designed a point cloud confidence sampling strategy based on a 3D Gaussian distribution to assign small confidence to the points with great error in depth estimation and filter them out according to the confidence.Then,we present a hierarchical geometric feature extraction module by aggregating the local neighborhood features and a dual transformer to capture the global geometric features in the point cloud.Finally,our detection framework is based on Point-Voxel-RCNN(PV-RCNN)with high-quality Pseudo-LiDAR and enriched geometric features as input.From the experimental results,our method achieves satisfactory results in monocular 3D object detection.展开更多
Monocular depth estimation is the basic task in computer vision.Its accuracy has tremendous improvement in the decade with the development of deep learning.However,the blurry boundary in the depth map is a serious pro...Monocular depth estimation is the basic task in computer vision.Its accuracy has tremendous improvement in the decade with the development of deep learning.However,the blurry boundary in the depth map is a serious problem.Researchers find that the blurry boundary is mainly caused by two factors.First,the low-level features,containing boundary and structure information,may be lost in deep networks during the convolution process.Second,themodel ignores the errors introduced by the boundary area due to the few portions of the boundary area in the whole area,during the backpropagation.Focusing on the factors mentioned above.Two countermeasures are proposed to mitigate the boundary blur problem.Firstly,we design a scene understanding module and scale transformmodule to build a lightweight fuse feature pyramid,which can deal with low-level feature loss effectively.Secondly,we propose a boundary-aware depth loss function to pay attention to the effects of the boundary’s depth value.Extensive experiments show that our method can predict the depth maps with clearer boundaries,and the performance of the depth accuracy based on NYU-Depth V2,SUN RGB-D,and iBims-1 are competitive.展开更多
To address the eccentric error of circular marks in camera calibration,a circle location method based on the invariance of collinear points and pole–polar constraint is proposed in this paper.Firstly,the centers of t...To address the eccentric error of circular marks in camera calibration,a circle location method based on the invariance of collinear points and pole–polar constraint is proposed in this paper.Firstly,the centers of the ellipses are extracted,and the real concentric circle center projection equation is established by exploiting the cross ratio invariance of the collinear points.Subsequently,since the infinite lines passing through the centers of the marks are parallel,the other center projection coordinates are expressed as the solution problem of linear equations.The problem of projection deviation caused by using the center of the ellipse as the real circle center projection is addressed,and the results are utilized as the true image points to achieve the high precision camera calibration.As demonstrated by the simulations and practical experiments,the proposed method performs a better location and calibration performance by achieving the actual center projection of circular marks.The relevant results confirm the precision and robustness of the proposed approach.展开更多
Traffic incident management (TIM) is a FHWA Every Day Counts initiative with the objective of reducing secondary crashes, improving travel reliability, and ensuring safety of responders. Agency roadside cameras play a...Traffic incident management (TIM) is a FHWA Every Day Counts initiative with the objective of reducing secondary crashes, improving travel reliability, and ensuring safety of responders. Agency roadside cameras play a critical role in TIM by helping dispatchers quickly identify the precise location of incidents when receiving reports from motorists with varying levels of spatial accuracy. Reconciling position reports that are often mile marker based, with cameras that operate in a Pan-Tilt-Zoom coordinate system relies on dispatchers having detailed knowledge for hundreds of cameras and perhaps some presets. During real-time incident dispatching, reducing the time it takes to identify the most relevant cameras and setting their view on the incident is an important opportunity to improve incident management dispatch times. This research develops a camera-to-mile marker mapping technique that automatically sets the camera view to a specified mile marker within the field-of-view of the camera. Over 350 traffic cameras along Indiana’s 2250 directional miles of interstate were mapped to approximately 5000 discrete locations that correspond to approximately 780 directional miles (~35% of interstate) of camera coverage. This newly developed technique will allow operators to quickly identify the nearest camera and set them to the reported location. This research also identifies segments on the interstate system with limited or no camera coverage for decision makers to prioritize future capital investments. This paper concludes with brief discussion on future research to automate the mapping using LiDAR data and to set the cameras after automatically detecting the events using connected vehicle trajectory data.展开更多
Monocular 6D pose estimation is a functional task in the field of com-puter vision and robotics.In recent years,2D-3D correspondence-based methods have achieved improved performance in multiview and depth data-based s...Monocular 6D pose estimation is a functional task in the field of com-puter vision and robotics.In recent years,2D-3D correspondence-based methods have achieved improved performance in multiview and depth data-based scenes.However,for monocular 6D pose estimation,these methods are affected by the prediction results of the 2D-3D correspondences and the robustness of the per-spective-n-point(PnP)algorithm.There is still a difference in the distance from the expected estimation effect.To obtain a more effective feature representation result,edge enhancement is proposed to increase the shape information of the object by analyzing the influence of inaccurate 2D-3D matching on 6D pose regression and comparing the effectiveness of the intermediate representation.Furthermore,although the transformation matrix is composed of rotation and translation matrices from 3D model points to 2D pixel points,the two variables are essentially different and the same network cannot be used for both variables in the regression process.Therefore,to improve the effectiveness of the PnP algo-rithm,this paper designs a dual-branch PnP network to predict rotation and trans-lation information.Finally,the proposed method is verified on the public LM,LM-O and YCB-Video datasets.The ADD(S)values of the proposed method are 94.2 and 62.84 on the LM and LM-O datasets,respectively.The AUC of ADD(-S)value on YCB-Video is 81.1.These experimental results show that the performance of the proposed method is superior to that of similar methods.展开更多
基金supported by the National Natural Science Foundation of China(Grant No.52272421)Shenzhen Fundamental Research Fund(Grant Number:JCYJ20190808142613246 and 20200803015912001).
文摘Estimating depth from images captured by camera sensors is crucial for the advancement of autonomous driving technologies and has gained significant attention in recent years.However,most previous methods rely on stacked pooling or stride convolution to extract high-level features,which can limit network performance and lead to information redundancy.This paper proposes an improved bidirectional feature pyramid module(BiFPN)and a channel attention module(Seblock:squeeze and excitation)to address these issues in existing methods based on monocular camera sensor.The Seblock redistributes channel feature weights to enhance useful information,while the improved BiFPN facilitates efficient fusion of multi-scale features.The proposed method is in an end-to-end solution without any additional post-processing,resulting in efficient depth estimation.Experiment results show that the proposed method is competitive with state-of-the-art algorithms and preserves fine-grained texture of scene depth.
基金supported in part by the Major Project for New Generation of AI (2018AAA0100400)the National Natural Science Foundation of China (61836014,U21B2042,62072457,62006231)the InnoHK Program。
文摘Monocular 3D object detection is challenging due to the lack of accurate depth information.Some methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input to augment the RGB images.Depth-based methods attempt to convert estimated depth maps to pseudo-LiDAR and then use LiDAR-based object detectors or focus on the perspective of image and depth fusion learning.However,they demonstrate limited performance and efficiency as a result of depth inaccuracy and complex fusion mode with convolutions.Different from these approaches,our proposed depth-guided vision transformer with a normalizing flows(NF-DVT)network uses normalizing flows to build priors in depth maps to achieve more accurate depth information.Then we develop a novel Swin-Transformer-based backbone with a fusion module to process RGB image patches and depth map patches with two separate branches and fuse them using cross-attention to exchange information with each other.Furthermore,with the help of pixel-wise relative depth values in depth maps,we develop new relative position embeddings in the cross-attention mechanism to capture more accurate sequence ordering of input tokens.Our method is the first Swin-Transformer-based backbone architecture for monocular 3D object detection.The experimental results on the KITTI and the challenging Waymo Open datasets show the effectiveness of our proposed method and superior performance over previous counterparts.
文摘Joint calibration of sensors is an important prerequisite in intelligent driving scene retrieval and recognition. A simple and efficient solution is proposed for solving the problem of automatic joint calibration registration between the monocular camera and the 16-line lidar. The study is divided into two parts: single-sensor independent calibration and multi-sensor joint registration, in which the selected objective world is used. The system associates the lidar coordinates with the camera coordinates. The lidar and the camera are used to obtain the normal vectors of the calibration plate and the point cloud data representing the calibration plate by the appropriate algorithm. Iterated closest points(ICP) is the method used for the iterative refinement of the registration.
基金supported by National Natural Science Foundation of China (Nos. 60874010 and 61070048)Innovation Program of Shanghai Municipal Education Commission (No. 11ZZ37)+1 种基金Fundamental Research Funds for the Central Universities (No. 009QJ12)Collaborative Construction Project of Beijing Municipal Commission of Education
文摘The rotation matrix estimation problem is a keypoint for mobile robot localization, navigation, and control. Based on the quaternion theory and the epipolar geometry, an extended Kalman filter (EKF) algorithm is proposed to estimate the rotation matrix by using a single-axis gyroscope and the image points correspondence from a monocular camera. The experimental results show that the precision of mobile robot s yaw angle estimated by the proposed EKF algorithm is much better than the results given by the image-only and gyroscope-only method, which demonstrates that our method is a preferable way to estimate the rotation for the autonomous mobile robot applications.
基金supported by the National Natural Science Foundation of China (No. 12220101005)Natural Science Foundation of Jiangsu Province (No. BK20220132)+2 种基金Primary Research and Development Plan of Jiangsu Province (No. BE2019002-3)Fundamental Research Funds for Central Universities (No. NG2022004)the Foundation of the Graduate Innovation Center in NUAA (No. xcxjh20210613)。
文摘A novel and fast three-dimensional reconstruction method for a Compton camera and its performance in radionuclide imaging is proposed and analyzed in this study. The conical surface sampling back-projection method with scattering angle correction(CSS-BP-SC) can quickly perform the back-projection process of the Compton cone and can be used to precompute the list-mode maximum likelihood expectation maximization(LM-MLEM). A dedicated parallel architecture was designed for the graphics processing unit acceleration of the back-projection and iteration stage of the CSS-BP-SC-based LM-MLEM. The imaging results of the two-point source Monte Carlo(MC) simulation demonstrate that by analyzing the full width at half maximum along the three coordinate axes, the CSS-BP-SC-based LM-MLEM can obtain imaging results comparable to those of the traditional reconstruction algorithm, that is, the simple back-projection-based LM-MLEM. The imaging results of the mouse phantom MC simulation and experiment demonstrate that the reconstruction results obtained by the proposed method sufficiently coincide with the set radioactivity distribution, and the speed increased by more than 664 times compared to the traditional reconstruction algorithm in the mouse phantom experiment. The proposed method will further advance the imaging applications of Compton cameras.
基金supported by Natural Science Foundation of Beijing Municipality (Beijing Natural Science Foundation)(No.7191005)。
文摘Compton camera-based prompt gamma(PG) imaging has been proposed for range verification during proton therapy. However, a deviation between the PG and dose distributions, as well as the difference between the reconstructed PG and exact values, limit the effectiveness of the approach in accurate range monitoring during clinical applications. The aim of the study was to realize a PG-based dose reconstruction with a Compton camera, thereby further improving the prediction accuracy of in vivo range verification and providing a novel method for beam monitoring during proton therapy. In this paper, we present an approach based on a subset-driven origin ensemble with resolution recovery and a double evolutionary algorithm to reconstruct the dose depth profile(DDP) from the gamma events obtained by a cadmium-zinc-telluride Compton camera with limited position and energy resolution. Simulations of proton pencil beams with clinical particle rate irradiating phantoms made of different materials and the CT-based thoracic phantom were used to evaluate the feasibility of the proposed method. The results show that for the monoenergetic proton pencil beam irradiating homogeneous-material box phantom,the accuracy of the reconstructed DDP was within 0.3 mm for range prediction and within 5.2% for dose prediction. In particular, for 1.6-Gy irradiation in the therapy simulation of thoracic tumors, the range deviation of the reconstructed spreadout Bragg peak was within 0.8 mm, and the relative dose deviation in the peak area was less than 7% compared to the exact values. The results demonstrate the potential and feasibility of the proposed method in future Compton-based accurate dose reconstruction and range verification during proton therapy.
基金This paper was supported in part by the National Science Founda-tion of China(Grant No.62103237).
文摘This paper proposes a person-following method based on monocular vision,which allows quadruped robots to track a target person in both indoor and outdoor environments with different illumination conditions.Our method is composed of a person detector,a Kalman filter(KF)tracker,and a re-identification module.To be more specific,the person detector uses a human pose estimation method to detect persons.The KF is applied to predict the position of the target person and update its state with detection results.A re-identification module is proposed to deal with distractions,where the Convolutional Channel Features(CCF)is used to extract appearance features and Online Boosting is used to distinguish the target person from others.Especially,we design a target recapture mechanism based on the Recurrent Neural Network(RNN).Combining motion information with appearance features,the system can accurately re-identify the target person.Without extra customized markers,our method can track the target person steadily in real-time only using a monocular camera.Experiments results can validate the robustness and effectiveness of the proposed method.
文摘Lanthanum bromide(LaBr_(3))crystal has a high energy resolution and time resolution and has been used in Compton cameras(CCs)over the past few decades.However,LaBr_(3) crystal arrays are difficult to process because LaBr_(3) is easy to crack and break;thus,few LaBr_(3)-based CC prototypes have been built.In this study,we designed and fabricated a large-pixel LaBr_(3) CC prototype and evaluated its performance with regard to position,energy,and angular resolution.We used two 10×10 LaBr_(3) crystal arrays with a pixel size of 5 mm×5 mm,silicon photomultipliers(SiPMs),and corresponding decoding circuits to construct our prototype.Additionally,a framework based on a Voronoi diagram and a lookup table was developed for list-mode projection data acquisition.Monte Carlo(MC)simulations based on Geant4 and experiments were conducted to evaluate the performance of our CC prototype.The lateral position resolution was 5 mm,and the maximum deviation in the depth direction was 2.5 and 5 mm for the scatterer and absorber,respectively.The corresponding measured energy resolu-tions were 7.65%and 8.44%,respectively,at 511 keV.The experimental results of ^(137)Cs point-like sources were consistent with the MC simulation results with regard to the spatial positions and full widths at half maximum(FWHMs).The angular resolution of the fabricated prototype was approximately 6°when a point-like ^(137)Cs source was centrally placed at a distance of 5 cm from the scatterer.We proposed and investigated a large-pixel LaBr_(3) CC for the first time and verified its feasibility for use in accurate spatial positioning of radiative sources with a high angular resolution.The proposed CC can satisfy the requirements of radiative source imaging and positioning in the nuclear industry and medical applications.
文摘Background Based on the seminal work proposed by Zhou et al., much of the recent progress in learning monocular visual odometry, i.e., depth and camera motion from monocular videos, can be attributed to the tricks in the training procedure, such as data augmentation and learning objectives. Methods Herein, we categorize a collection of such tricks through the theoretical examination and empirical evaluation of their effects on the final accuracy of the visual odometry. Results/Conclusions By combining the aforementioned tricks, we were able to significantly improve a baseline model adapted from SfMLearner without additional inference costs. Furthermore, we analyzed the principles of these tricks and the reason for their success. Practical guidelines for future research are also presented.
基金National Natural Science Foundation of China(11805269)West Light Talent Training Plan of the Chinese Academy of Sciences(2022-XBQNXZ-010)Science and Technology Innovation Leading Talent Project of Xinjiang Uygur Autonomous Region(2022TSYCLJ0042)。
文摘Theγ-rays are widely and abundantly present in strong nuclear radiation environments,and when they act on the camera equipment used to obtain environmental visual information on nuclear robots,radiation effects will occur,which will degrade the performance of the camera system,reduce the imaging quality,and even cause catastrophic consequences.Color reducibility is an important index for evaluating the imaging quality of color camera,but its degradation mechanism in a nuclear radiation environment is still unclear.In this paper,theγ-ray irradiation experiments of CMOS cameras were carried out to analyse the degradation law of the camera’s color reducibility with cumulative irradiation and reveal the degradation mechanism of the color information of the CMOS camera underγ-ray irradiation.The results show that the spectral response of CMOS image sensor(CIS)and the spectral transmittance of lens after irradiation affect the values of a^(*)and b^(*)in the LAB color model.While the full well capacity(FWC)of CIS and transmittance of lens affect the value of L^(*)in the LAB color model,thus increase color difference and reduce brightness,the combined effect of color difference and brightness degradation will reduce the color reducibility of CMOS cameras.Therefore,the degradation of the color information of the CMOS camera afterγ-ray irradiation mainly comes from the changes in the FWC and spectral response of CIS,and the spectral transmittance of lens.
基金supported by the National Key Research and Development Program of China(2020YFB1807500)the National Natural Science Foundation of China(62072360,62001357,62172438,61901367)+4 种基金the key research and development plan of Shaanxi province(2021ZDLGY02-09,2023-GHZD-44,2023-ZDLGY-54)the Natural Science Foundation of Guangdong Province of China(2022A1515010988)Key Project on Artificial Intelligence of Xi'an Science and Technology Plan(2022JH-RGZN-0003,2022JH-RGZN-0103,2022JH-CLCJ-0053)Xi'an Science and Technology Plan(20RGZN0005)the Proof-ofconcept fund from Hangzhou Research Institute of Xidian University(GNYZ2023QC0201).
文摘The high bandwidth and low latency of 6G network technology enable the successful application of monocular 3D object detection on vehicle platforms.Monocular 3D-object-detection-based Pseudo-LiDAR is a low-cost,lowpower solution compared to LiDAR solutions in the field of autonomous driving.However,this technique has some problems,i.e.,(1)the poor quality of generated Pseudo-LiDAR point clouds resulting from the nonlinear error distribution of monocular depth estimation and(2)the weak representation capability of point cloud features due to the neglected global geometric structure features of point clouds existing in LiDAR-based 3D detection networks.Therefore,we proposed a Pseudo-LiDAR confidence sampling strategy and a hierarchical geometric feature extraction module for monocular 3D object detection.We first designed a point cloud confidence sampling strategy based on a 3D Gaussian distribution to assign small confidence to the points with great error in depth estimation and filter them out according to the confidence.Then,we present a hierarchical geometric feature extraction module by aggregating the local neighborhood features and a dual transformer to capture the global geometric features in the point cloud.Finally,our detection framework is based on Point-Voxel-RCNN(PV-RCNN)with high-quality Pseudo-LiDAR and enriched geometric features as input.From the experimental results,our method achieves satisfactory results in monocular 3D object detection.
基金supported in part by School Research Projects of Wuyi University (No.5041700175).
文摘Monocular depth estimation is the basic task in computer vision.Its accuracy has tremendous improvement in the decade with the development of deep learning.However,the blurry boundary in the depth map is a serious problem.Researchers find that the blurry boundary is mainly caused by two factors.First,the low-level features,containing boundary and structure information,may be lost in deep networks during the convolution process.Second,themodel ignores the errors introduced by the boundary area due to the few portions of the boundary area in the whole area,during the backpropagation.Focusing on the factors mentioned above.Two countermeasures are proposed to mitigate the boundary blur problem.Firstly,we design a scene understanding module and scale transformmodule to build a lightweight fuse feature pyramid,which can deal with low-level feature loss effectively.Secondly,we propose a boundary-aware depth loss function to pay attention to the effects of the boundary’s depth value.Extensive experiments show that our method can predict the depth maps with clearer boundaries,and the performance of the depth accuracy based on NYU-Depth V2,SUN RGB-D,and iBims-1 are competitive.
基金supported by the Aerospace Science and Technology Joint Fund(6141B061505)the National Natural Science Foundation of China(61473100).
文摘To address the eccentric error of circular marks in camera calibration,a circle location method based on the invariance of collinear points and pole–polar constraint is proposed in this paper.Firstly,the centers of the ellipses are extracted,and the real concentric circle center projection equation is established by exploiting the cross ratio invariance of the collinear points.Subsequently,since the infinite lines passing through the centers of the marks are parallel,the other center projection coordinates are expressed as the solution problem of linear equations.The problem of projection deviation caused by using the center of the ellipse as the real circle center projection is addressed,and the results are utilized as the true image points to achieve the high precision camera calibration.As demonstrated by the simulations and practical experiments,the proposed method performs a better location and calibration performance by achieving the actual center projection of circular marks.The relevant results confirm the precision and robustness of the proposed approach.
文摘Traffic incident management (TIM) is a FHWA Every Day Counts initiative with the objective of reducing secondary crashes, improving travel reliability, and ensuring safety of responders. Agency roadside cameras play a critical role in TIM by helping dispatchers quickly identify the precise location of incidents when receiving reports from motorists with varying levels of spatial accuracy. Reconciling position reports that are often mile marker based, with cameras that operate in a Pan-Tilt-Zoom coordinate system relies on dispatchers having detailed knowledge for hundreds of cameras and perhaps some presets. During real-time incident dispatching, reducing the time it takes to identify the most relevant cameras and setting their view on the incident is an important opportunity to improve incident management dispatch times. This research develops a camera-to-mile marker mapping technique that automatically sets the camera view to a specified mile marker within the field-of-view of the camera. Over 350 traffic cameras along Indiana’s 2250 directional miles of interstate were mapped to approximately 5000 discrete locations that correspond to approximately 780 directional miles (~35% of interstate) of camera coverage. This newly developed technique will allow operators to quickly identify the nearest camera and set them to the reported location. This research also identifies segments on the interstate system with limited or no camera coverage for decision makers to prioritize future capital investments. This paper concludes with brief discussion on future research to automate the mapping using LiDAR data and to set the cameras after automatically detecting the events using connected vehicle trajectory data.
基金This work was supported by the National Natural Science Foundation of China(No.61871196 and 62001176)the Natural Science Foundation of Fujian Province of China(No.2019J01082 and 2020J01085)the Promotion Program for Young and Middle-aged Teachers in Science and Technology Research of Huaqiao University(ZQN-YX601).
文摘Monocular 6D pose estimation is a functional task in the field of com-puter vision and robotics.In recent years,2D-3D correspondence-based methods have achieved improved performance in multiview and depth data-based scenes.However,for monocular 6D pose estimation,these methods are affected by the prediction results of the 2D-3D correspondences and the robustness of the per-spective-n-point(PnP)algorithm.There is still a difference in the distance from the expected estimation effect.To obtain a more effective feature representation result,edge enhancement is proposed to increase the shape information of the object by analyzing the influence of inaccurate 2D-3D matching on 6D pose regression and comparing the effectiveness of the intermediate representation.Furthermore,although the transformation matrix is composed of rotation and translation matrices from 3D model points to 2D pixel points,the two variables are essentially different and the same network cannot be used for both variables in the regression process.Therefore,to improve the effectiveness of the PnP algo-rithm,this paper designs a dual-branch PnP network to predict rotation and trans-lation information.Finally,the proposed method is verified on the public LM,LM-O and YCB-Video datasets.The ADD(S)values of the proposed method are 94.2 and 62.84 on the LM and LM-O datasets,respectively.The AUC of ADD(-S)value on YCB-Video is 81.1.These experimental results show that the performance of the proposed method is superior to that of similar methods.
