Lidar and visual data are affected heavily in adverse weather conditions due to sensing mechanisms,which bring potential safety hazards for vehicle navigation.Radar sensing is desirable to build a more robust navigati...Lidar and visual data are affected heavily in adverse weather conditions due to sensing mechanisms,which bring potential safety hazards for vehicle navigation.Radar sensing is desirable to build a more robust navigation system.In this paper,a cross-modality radar localisation on prior lidar maps is presented.Specifically,the proposed workflow consists of two parts:first,bird's-eye-view radar images are transferred to fake lidar images by training a generative adversarial network offline.Then with online radar scans,a Monte Carlo localisation framework is built to track the robot pose on lidar maps.The whole online localisation system only needs a rotating radar sensor and a pre-built global lidar map.In the experimental section,the authors conduct an ablation study on image settings and test the proposed system on Oxford Radar Robot Car Dataset.The promising results show that the proposed localisation system could track the robot pose successfully,thus demonstrating the feasibility of radar style transfer for metric robot localisation on lidar maps.展开更多
In this paper,the adaptive lifting scheme (ALS) and local gradient maps (LGM) are proposed to isolate the transient feature components from the gearbox vibration signals. Based on entropy minimization rule,the ALS is ...In this paper,the adaptive lifting scheme (ALS) and local gradient maps (LGM) are proposed to isolate the transient feature components from the gearbox vibration signals. Based on entropy minimization rule,the ALS is employed to change properties of an initial wavelet and design adaptive wavelet. Then LGM is applied to characterize the transient feature components in detail signal of decomposition results using ALS. In the present studies, the orthogonal Daubechies 4 (Db 4) wavelet is used as the initial wavelet. The proposed method is applied to both simulated signals and vibration signals acquired from a gearbox for periodic impulses detection. The two conventional methods (cepstrum analysis and Hilbert envelope analysis) and the orthogonal Db4 wavelet are also used to analyze the same signals for comparison. The results demonstrate that the proposed method is more effective in extracting transient components from noisy signals.展开更多
An extended Kalman filter approach of simultaneous localization and mapping(SLAM) was proposed based on local maps. A local frame of reference was established periodically at the position of the robot, and then the ob...An extended Kalman filter approach of simultaneous localization and mapping(SLAM) was proposed based on local maps. A local frame of reference was established periodically at the position of the robot, and then the observations of the robot and landmarks were fused into the global frame of reference. Because of the independence of the local map, the approach does not cumulate the estimate and calculation errors which are produced by SLAM using Kalman filter directly. At the same time, it reduces the computational complexity. This method is proven correct and feasible in simulation experiments.展开更多
The processing maps were used to identify the optimal forging parameters of Ti-24A1- 17Nb-0.5Mo alloy by evaluating the flow data according to the DMM model. The actual local strain rate and strain distribution in the...The processing maps were used to identify the optimal forging parameters of Ti-24A1- 17Nb-0.5Mo alloy by evaluating the flow data according to the DMM model. The actual local strain rate and strain distribution in the samples were obtained by finite element calculations. The local microstructures of the deformed samples were related to the local deformation parameters and correlated with the processing maps at 0.3, 0.4, 0.5 and 0.6 of logarithmic strain. Flow regimes predicted by DMM analysis were then correlated with the local microstructural observations. Five domains of efficient coefficient could be distinguished. Unstable regions were microstructurally related to shear band formation within the （~2~B2 phase deformation field, and to flow localiza- tion at grain boundaries of B2 phase in the near B2 phase deformation field. Stable flow regimes were shown to be associated with dynamic globularization of the plate- like a2 in the a2＋B2 phase deformation zone, and with dynamic recrystallization of B2 in the near B2 phase deformation zone.展开更多
Digital maps of soil properties are now widely available.End-users now can access several digital soil mapping(DSM)products of soil properties,produced using different models,calibration/training data,and covariates a...Digital maps of soil properties are now widely available.End-users now can access several digital soil mapping(DSM)products of soil properties,produced using different models,calibration/training data,and covariates at various spatial scales from global to local.Therefore,there is an urgent need to provide easy-to-understand tools to communicate map uncertainty and help end-users assess the reliability of DSM products for use at local scales.In this study,we used a large amount of hand-feel soil texture(HFST)data to assess the performance of various published DSM products on the prediction of soil particle size distribution in Central France.We tested four DSM products for soil texture prediction developed at various scales(global,continental,national,and regional)by comparing their predictions with approximately 3200 HFST observations realized on a 1:50000 soil survey conducted after release of these DSM products.We used both visual comparisons and quantitative indicators to match the DSM predictions and HFST observations.The comparison between the low-cost HFST observations and DSM predictions clearly showed the applicability of various DSM products,with the prediction accuracy increasing from global to regional predictions.This simple evaluation can determine which products can be used at the local scale and if more accurate DSM products are required.展开更多
A new methodology of comparing digital raster maps was proposed which allows not only detecting changes in the maps, but also obtaining quantitative measures of the importance of selected differences. Procedure of obj...A new methodology of comparing digital raster maps was proposed which allows not only detecting changes in the maps, but also obtaining quantitative measures of the importance of selected differences. Procedure of object interpretation of satellite images and forming of OMT (Object Map of Territory) is described. A list of allowable differences between two OMTs is defined. Two steps technique of quantitative measuring is proposed. At the first stage functions are constructed for calculating local measures of differences in the amount, areas and locations of objects on the map, as well as relations between the objects. In the second stage local measures are used to calculate the integral measure in order to get generalized assessment of difference between maps. The methods for constructing functions which calculate local and integral measures of differences are described. Examples of comparing and measuring the differences between OMTs are provided. Obtained results by utilizing this technique can be used to analyze trends, forecast of development and might be helpful for choosing most efficient scenarios for sustainable spatial planning and land management.展开更多
Ecological monitoring vehicles are equipped with a range of sensors and monitoring devices designed to gather data on ecological and environmental factors.These vehicles are crucial in various fields,including environ...Ecological monitoring vehicles are equipped with a range of sensors and monitoring devices designed to gather data on ecological and environmental factors.These vehicles are crucial in various fields,including environmental science research,ecological and environmental monitoring projects,disaster response,and emergency management.A key method employed in these vehicles for achieving high-precision positioning is LiDAR(lightlaser detection and ranging)-Visual Simultaneous Localization and Mapping(SLAM).However,maintaining highprecision localization in complex scenarios,such as degraded environments or when dynamic objects are present,remains a significant challenge.To address this issue,we integrate both semantic and texture information from LiDAR and cameras to enhance the robustness and efficiency of data registration.Specifically,semantic information simplifies the modeling of scene elements,reducing the reliance on dense point clouds,which can be less efficient.Meanwhile,visual texture information complements LiDAR-Visual localization by providing additional contextual details.By incorporating semantic and texture details frompaired images and point clouds,we significantly improve the quality of data association,thereby increasing the success rate of localization.This approach not only enhances the operational capabilities of ecological monitoring vehicles in complex environments but also contributes to improving the overall efficiency and effectiveness of ecological monitoring and environmental protection efforts.展开更多
The concept of local shock strength and a quantitative measure index str of local shock strength are proposed,derived from the oblique shock relation and the monotonic relationship between total pressure loss ratio an...The concept of local shock strength and a quantitative measure index str of local shock strength are proposed,derived from the oblique shock relation and the monotonic relationship between total pressure loss ratio and normal Mach number.Utilizing the high density gradient characteristic of shock waves and the oblique shock relation,a post-processing algorithm for two-dimensional flow field data is developed.The objective of the post-processing algorithm is to obtain specific shock wave location coordinates and calculate the corresponding str from flow filed data under the calibration of the oblique shock relation.Valida-tion of this post-processing algorithm is conducted using a standard model example that can be solved analytically.Combining the concept of local shock strength with the post-processing algorithm,a local shock strength quantitative mapping approach is established for the first time.This approach enables a quantitative measure and visualization of local shock strength at distinct locations,represented by color mapping on the shock structures.The approach can be applied to post-processing numerical sim-ulation data of two-dimensional flows.Applications to the intersection of two left-running oblique shock waves(straight shock waves),the bow shock in front of a cylinder(curved shock wave),and Mach reflection(mixed straight and curved shock waves)demonstrate the accuracy,and effectiveness of the mapping approach in investigating diverse shock wave phenomena.The quan-titative mapping approach of str may be a valuable tool in the design of supersonic/hypersonic vehicles and the exploration of shock wave evolution.展开更多
The Grey Wolf Optimization(GWO)algorithm is acknowledged as an effective method for rock acoustic emission localization.However,the conventional GWO algorithm encounters challenges related to solution accuracy and con...The Grey Wolf Optimization(GWO)algorithm is acknowledged as an effective method for rock acoustic emission localization.However,the conventional GWO algorithm encounters challenges related to solution accuracy and convergence speed.To address these concerns,this paper develops a Simplex Improved Grey Wolf Optimizer(SMIGWO)algorithm.The randomly generating initial populations are replaced with the iterative chaotic sequences.The search process is optimized using the convergence factor optimization algorithm based on the inverse incompleteГfunction.The simplex method is utilized to address issues related to poorly positioned grey wolves.Experimental results demonstrate that,compared to the conventional GWO algorithm-based AE localization algorithm,the proposed algorithm achieves a higher solution accuracy and showcases a shorter search time.Additionally,the algorithm demonstrates fewer convergence steps,indicating superior convergence efficiency.These findings highlight that the proposed SMIGWO algorithm offers enhanced solution accuracy,stability,and optimization performance.The benefits of the SMIGWO algorithm extend universally across various materials,such as aluminum,granite,and sandstone,showcasing consistent effectiveness irrespective of material type.Consequently,this algorithm emerges as a highly effective tool for identifying acoustic emission signals and improving the precision of rock acoustic emission localization.展开更多
Real-time indoor camera localization is a significant problem in indoor robot navigation and surveillance systems.The scene can change during the image sequence and plays a vital role in the localization performance o...Real-time indoor camera localization is a significant problem in indoor robot navigation and surveillance systems.The scene can change during the image sequence and plays a vital role in the localization performance of robotic applications in terms of accuracy and speed.This research proposed a real-time indoor camera localization system based on a recurrent neural network that detects scene change during the image sequence.An annotated image dataset trains the proposed system and predicts the camera pose in real-time.The system mainly improved the localization performance of indoor cameras by more accurately predicting the camera pose.It also recognizes the scene changes during the sequence and evaluates the effects of these changes.This system achieved high accuracy and real-time performance.The scene change detection process was performed using visual rhythm and the proposed recurrent deep architecture,which performed camera pose prediction and scene change impact evaluation.Overall,this study proposed a novel real-time localization system for indoor cameras that detects scene changes and shows how they affect localization performance.展开更多
在智慧矿山建设的背景下,智能化设备的应用日益成为矿山智慧化改造的主要内容,用于巡检、危险区域勘测等任务的煤矿井下智能机器人运行依赖于数字地图构建和机器人自身定位,但大多数传统的定位方法在煤矿井下出现了低效甚至失效的情况,...在智慧矿山建设的背景下,智能化设备的应用日益成为矿山智慧化改造的主要内容,用于巡检、危险区域勘测等任务的煤矿井下智能机器人运行依赖于数字地图构建和机器人自身定位,但大多数传统的定位方法在煤矿井下出现了低效甚至失效的情况,同步定位与建图技术(Simultaneous Localization and Mapping,SLAM)成为了煤矿井下智能机器人定位方法的较优选择。然而,受制于激光雷达的高成本,以及相机在井下的低光照环境性能不佳,需要设计一种兼顾低成本和具有井下低光照环境适应性的SLAM定位方法,故提出了一种具有井下暗光照适应性煤矿井下机器人定位方法。首先,采集了陕西省宝鸡市凤县某煤矿井下的实景图像和SLAM所需的相机与IMU数据,根据图像制作了非匹配的暗光与正常光数据集,经过数据扩增达到3560张图像。设计了结合自注意力模块的EnlightenGAN图像增强网络,在不依赖配对数据集的情况下兼顾图像不同区域的依赖关系应对图像光照不均区域。在ORB-SLAM3框架的基础上,引入全局部图像检测对输入图像进行筛分,引入基于解析解的IMU初始化改进策略提高初始化速度,并引入了改进的图像增强网络对低光照以及光照不均的图像进行增强处理。在EuRoC数据集上的试验表明,基于图像增强的煤矿井下智能机器人定位方法能够在低光照环境下降低13.7%的ERMS和15.24%的ESD。在2个实际煤矿巷道场景中,系统能够识别低光照环境、增加SLAM系统提取的特征点数量,减少定位轨迹的漂移现象,最终改善系统在巷道低光照区域的定位效果。展开更多
煤矿井下视觉同步定位与地图构建SLAM(Simultaneous Localization and Mapping)应用中,光照变化与低纹理场景严重影响特征点的提取和匹配结果,导致位姿估计失败,影响定位精度。提出一种基于改进定向快速旋转二值描述符ORB(Oriented Fast...煤矿井下视觉同步定位与地图构建SLAM(Simultaneous Localization and Mapping)应用中,光照变化与低纹理场景严重影响特征点的提取和匹配结果,导致位姿估计失败,影响定位精度。提出一种基于改进定向快速旋转二值描述符ORB(Oriented Fast and Rotated Brief)-SLAM3算法的煤矿井下移动机器人双目视觉定位算法SL-SLAM。针对光照变化场景,在前端使用光照稳定性的Super-Point特征点提取网络替换原始ORB特征点提取算法,并提出一种特征点网格限定法,有效剔除无效特征点区域,增加位姿估计稳定性。针对低纹理场景,在前端引入稳定的线段检测器LSD(Line Segment Detector)线特征提取算法,并提出一种点线联合算法,按照特征点网格对线特征进行分组,根据特征点的匹配结果进行线特征匹配,降低线特征匹配复杂度,节约位姿估计时间。构建了点特征和线特征的重投影误差模型,在线特征残差模型中添加角度约束,通过点特征和线特征的位姿增量雅可比矩阵建立点线特征重投影误差统一成本函数。局部建图线程使用ORB-SLAM3经典的局部优化方法调整点、线特征和关键帧位姿,并在后端线程中进行回环修正、子图融合和全局捆绑调整BA(Bundle Adjustment)。在EuRoC数据集上的试验结果表明,SL-SLAM的绝对位姿误差APE(Absolute Pose Error)指标优于其他对比算法,并取得了与真值最接近的轨迹预测结果:均方根误差相较于ORB-SLAM3降低了17.3%。在煤矿井下模拟场景中的试验结果表明,SL-SLAM能适应光照变化和低纹理场景,可以满足煤矿井下移动机器人的定位精度和稳定性要求。展开更多
基金National Key R&D Program of China,Grant/Award Number:2020YFB1313300National Nature Science Foundation of China under Grant,Grant/Award Number:61903332Hong Kong Center for Construction Robotics(InnoHK center supported by Hong Kong ITC)。
文摘Lidar and visual data are affected heavily in adverse weather conditions due to sensing mechanisms,which bring potential safety hazards for vehicle navigation.Radar sensing is desirable to build a more robust navigation system.In this paper,a cross-modality radar localisation on prior lidar maps is presented.Specifically,the proposed workflow consists of two parts:first,bird's-eye-view radar images are transferred to fake lidar images by training a generative adversarial network offline.Then with online radar scans,a Monte Carlo localisation framework is built to track the robot pose on lidar maps.The whole online localisation system only needs a rotating radar sensor and a pre-built global lidar map.In the experimental section,the authors conduct an ablation study on image settings and test the proposed system on Oxford Radar Robot Car Dataset.The promising results show that the proposed localisation system could track the robot pose successfully,thus demonstrating the feasibility of radar style transfer for metric robot localisation on lidar maps.
基金Higher School Specialized Research Fund for the Doctoral Program Funding Issue(No.2011021120032)Fundamental Research Funds for the Central Universities(No.2012jdhz23)
文摘In this paper,the adaptive lifting scheme (ALS) and local gradient maps (LGM) are proposed to isolate the transient feature components from the gearbox vibration signals. Based on entropy minimization rule,the ALS is employed to change properties of an initial wavelet and design adaptive wavelet. Then LGM is applied to characterize the transient feature components in detail signal of decomposition results using ALS. In the present studies, the orthogonal Daubechies 4 (Db 4) wavelet is used as the initial wavelet. The proposed method is applied to both simulated signals and vibration signals acquired from a gearbox for periodic impulses detection. The two conventional methods (cepstrum analysis and Hilbert envelope analysis) and the orthogonal Db4 wavelet are also used to analyze the same signals for comparison. The results demonstrate that the proposed method is more effective in extracting transient components from noisy signals.
基金Project(60234030) supported by the National Natural Science Foundation of China project(A1420060159) supported by the National Basic Research
文摘An extended Kalman filter approach of simultaneous localization and mapping(SLAM) was proposed based on local maps. A local frame of reference was established periodically at the position of the robot, and then the observations of the robot and landmarks were fused into the global frame of reference. Because of the independence of the local map, the approach does not cumulate the estimate and calculation errors which are produced by SLAM using Kalman filter directly. At the same time, it reduces the computational complexity. This method is proven correct and feasible in simulation experiments.
文摘The processing maps were used to identify the optimal forging parameters of Ti-24A1- 17Nb-0.5Mo alloy by evaluating the flow data according to the DMM model. The actual local strain rate and strain distribution in the samples were obtained by finite element calculations. The local microstructures of the deformed samples were related to the local deformation parameters and correlated with the processing maps at 0.3, 0.4, 0.5 and 0.6 of logarithmic strain. Flow regimes predicted by DMM analysis were then correlated with the local microstructural observations. Five domains of efficient coefficient could be distinguished. Unstable regions were microstructurally related to shear band formation within the （~2~B2 phase deformation field, and to flow localiza- tion at grain boundaries of B2 phase in the near B2 phase deformation field. Stable flow regimes were shown to be associated with dynamic globularization of the plate- like a2 in the a2＋B2 phase deformation zone, and with dynamic recrystallization of B2 in the near B2 phase deformation zone.
文摘Digital maps of soil properties are now widely available.End-users now can access several digital soil mapping(DSM)products of soil properties,produced using different models,calibration/training data,and covariates at various spatial scales from global to local.Therefore,there is an urgent need to provide easy-to-understand tools to communicate map uncertainty and help end-users assess the reliability of DSM products for use at local scales.In this study,we used a large amount of hand-feel soil texture(HFST)data to assess the performance of various published DSM products on the prediction of soil particle size distribution in Central France.We tested four DSM products for soil texture prediction developed at various scales(global,continental,national,and regional)by comparing their predictions with approximately 3200 HFST observations realized on a 1:50000 soil survey conducted after release of these DSM products.We used both visual comparisons and quantitative indicators to match the DSM predictions and HFST observations.The comparison between the low-cost HFST observations and DSM predictions clearly showed the applicability of various DSM products,with the prediction accuracy increasing from global to regional predictions.This simple evaluation can determine which products can be used at the local scale and if more accurate DSM products are required.
文摘A new methodology of comparing digital raster maps was proposed which allows not only detecting changes in the maps, but also obtaining quantitative measures of the importance of selected differences. Procedure of object interpretation of satellite images and forming of OMT (Object Map of Territory) is described. A list of allowable differences between two OMTs is defined. Two steps technique of quantitative measuring is proposed. At the first stage functions are constructed for calculating local measures of differences in the amount, areas and locations of objects on the map, as well as relations between the objects. In the second stage local measures are used to calculate the integral measure in order to get generalized assessment of difference between maps. The methods for constructing functions which calculate local and integral measures of differences are described. Examples of comparing and measuring the differences between OMTs are provided. Obtained results by utilizing this technique can be used to analyze trends, forecast of development and might be helpful for choosing most efficient scenarios for sustainable spatial planning and land management.
基金supported by the project“GEF9874:Strengthening Coordinated Approaches to Reduce Invasive Alien Species(lAS)Threats to Globally Significant Agrobiodiversity and Agroecosystems in China”funding from the Excellent Talent Training Funding Project in Dongcheng District,Beijing,with project number 2024-dchrcpyzz-9.
文摘Ecological monitoring vehicles are equipped with a range of sensors and monitoring devices designed to gather data on ecological and environmental factors.These vehicles are crucial in various fields,including environmental science research,ecological and environmental monitoring projects,disaster response,and emergency management.A key method employed in these vehicles for achieving high-precision positioning is LiDAR(lightlaser detection and ranging)-Visual Simultaneous Localization and Mapping(SLAM).However,maintaining highprecision localization in complex scenarios,such as degraded environments or when dynamic objects are present,remains a significant challenge.To address this issue,we integrate both semantic and texture information from LiDAR and cameras to enhance the robustness and efficiency of data registration.Specifically,semantic information simplifies the modeling of scene elements,reducing the reliance on dense point clouds,which can be less efficient.Meanwhile,visual texture information complements LiDAR-Visual localization by providing additional contextual details.By incorporating semantic and texture details frompaired images and point clouds,we significantly improve the quality of data association,thereby increasing the success rate of localization.This approach not only enhances the operational capabilities of ecological monitoring vehicles in complex environments but also contributes to improving the overall efficiency and effectiveness of ecological monitoring and environmental protection efforts.
基金supported by the National Natural Science Foundation of China(Grant No.12372233)the Fund of NPU-Duke China Seed Program(Grant No.119003067)the“111 Project”of China(Grant No.B17037-106).
文摘The concept of local shock strength and a quantitative measure index str of local shock strength are proposed,derived from the oblique shock relation and the monotonic relationship between total pressure loss ratio and normal Mach number.Utilizing the high density gradient characteristic of shock waves and the oblique shock relation,a post-processing algorithm for two-dimensional flow field data is developed.The objective of the post-processing algorithm is to obtain specific shock wave location coordinates and calculate the corresponding str from flow filed data under the calibration of the oblique shock relation.Valida-tion of this post-processing algorithm is conducted using a standard model example that can be solved analytically.Combining the concept of local shock strength with the post-processing algorithm,a local shock strength quantitative mapping approach is established for the first time.This approach enables a quantitative measure and visualization of local shock strength at distinct locations,represented by color mapping on the shock structures.The approach can be applied to post-processing numerical sim-ulation data of two-dimensional flows.Applications to the intersection of two left-running oblique shock waves(straight shock waves),the bow shock in front of a cylinder(curved shock wave),and Mach reflection(mixed straight and curved shock waves)demonstrate the accuracy,and effectiveness of the mapping approach in investigating diverse shock wave phenomena.The quan-titative mapping approach of str may be a valuable tool in the design of supersonic/hypersonic vehicles and the exploration of shock wave evolution.
基金support from the National Science Foundation of China(52304137,5192780752274124,52325403)Tiandi Science and Technology Co.,Ltd.(2022-2-TDMS012 and SKLIS202417)Sichuan University(SKHL2215).
文摘The Grey Wolf Optimization(GWO)algorithm is acknowledged as an effective method for rock acoustic emission localization.However,the conventional GWO algorithm encounters challenges related to solution accuracy and convergence speed.To address these concerns,this paper develops a Simplex Improved Grey Wolf Optimizer(SMIGWO)algorithm.The randomly generating initial populations are replaced with the iterative chaotic sequences.The search process is optimized using the convergence factor optimization algorithm based on the inverse incompleteГfunction.The simplex method is utilized to address issues related to poorly positioned grey wolves.Experimental results demonstrate that,compared to the conventional GWO algorithm-based AE localization algorithm,the proposed algorithm achieves a higher solution accuracy and showcases a shorter search time.Additionally,the algorithm demonstrates fewer convergence steps,indicating superior convergence efficiency.These findings highlight that the proposed SMIGWO algorithm offers enhanced solution accuracy,stability,and optimization performance.The benefits of the SMIGWO algorithm extend universally across various materials,such as aluminum,granite,and sandstone,showcasing consistent effectiveness irrespective of material type.Consequently,this algorithm emerges as a highly effective tool for identifying acoustic emission signals and improving the precision of rock acoustic emission localization.
文摘Real-time indoor camera localization is a significant problem in indoor robot navigation and surveillance systems.The scene can change during the image sequence and plays a vital role in the localization performance of robotic applications in terms of accuracy and speed.This research proposed a real-time indoor camera localization system based on a recurrent neural network that detects scene change during the image sequence.An annotated image dataset trains the proposed system and predicts the camera pose in real-time.The system mainly improved the localization performance of indoor cameras by more accurately predicting the camera pose.It also recognizes the scene changes during the sequence and evaluates the effects of these changes.This system achieved high accuracy and real-time performance.The scene change detection process was performed using visual rhythm and the proposed recurrent deep architecture,which performed camera pose prediction and scene change impact evaluation.Overall,this study proposed a novel real-time localization system for indoor cameras that detects scene changes and shows how they affect localization performance.
文摘在智慧矿山建设的背景下,智能化设备的应用日益成为矿山智慧化改造的主要内容,用于巡检、危险区域勘测等任务的煤矿井下智能机器人运行依赖于数字地图构建和机器人自身定位,但大多数传统的定位方法在煤矿井下出现了低效甚至失效的情况,同步定位与建图技术(Simultaneous Localization and Mapping,SLAM)成为了煤矿井下智能机器人定位方法的较优选择。然而,受制于激光雷达的高成本,以及相机在井下的低光照环境性能不佳,需要设计一种兼顾低成本和具有井下低光照环境适应性的SLAM定位方法,故提出了一种具有井下暗光照适应性煤矿井下机器人定位方法。首先,采集了陕西省宝鸡市凤县某煤矿井下的实景图像和SLAM所需的相机与IMU数据,根据图像制作了非匹配的暗光与正常光数据集,经过数据扩增达到3560张图像。设计了结合自注意力模块的EnlightenGAN图像增强网络,在不依赖配对数据集的情况下兼顾图像不同区域的依赖关系应对图像光照不均区域。在ORB-SLAM3框架的基础上,引入全局部图像检测对输入图像进行筛分,引入基于解析解的IMU初始化改进策略提高初始化速度,并引入了改进的图像增强网络对低光照以及光照不均的图像进行增强处理。在EuRoC数据集上的试验表明,基于图像增强的煤矿井下智能机器人定位方法能够在低光照环境下降低13.7%的ERMS和15.24%的ESD。在2个实际煤矿巷道场景中,系统能够识别低光照环境、增加SLAM系统提取的特征点数量,减少定位轨迹的漂移现象,最终改善系统在巷道低光照区域的定位效果。
文摘煤矿井下视觉同步定位与地图构建SLAM(Simultaneous Localization and Mapping)应用中,光照变化与低纹理场景严重影响特征点的提取和匹配结果,导致位姿估计失败,影响定位精度。提出一种基于改进定向快速旋转二值描述符ORB(Oriented Fast and Rotated Brief)-SLAM3算法的煤矿井下移动机器人双目视觉定位算法SL-SLAM。针对光照变化场景,在前端使用光照稳定性的Super-Point特征点提取网络替换原始ORB特征点提取算法,并提出一种特征点网格限定法,有效剔除无效特征点区域,增加位姿估计稳定性。针对低纹理场景,在前端引入稳定的线段检测器LSD(Line Segment Detector)线特征提取算法,并提出一种点线联合算法,按照特征点网格对线特征进行分组,根据特征点的匹配结果进行线特征匹配,降低线特征匹配复杂度,节约位姿估计时间。构建了点特征和线特征的重投影误差模型,在线特征残差模型中添加角度约束,通过点特征和线特征的位姿增量雅可比矩阵建立点线特征重投影误差统一成本函数。局部建图线程使用ORB-SLAM3经典的局部优化方法调整点、线特征和关键帧位姿,并在后端线程中进行回环修正、子图融合和全局捆绑调整BA(Bundle Adjustment)。在EuRoC数据集上的试验结果表明,SL-SLAM的绝对位姿误差APE(Absolute Pose Error)指标优于其他对比算法,并取得了与真值最接近的轨迹预测结果:均方根误差相较于ORB-SLAM3降低了17.3%。在煤矿井下模拟场景中的试验结果表明,SL-SLAM能适应光照变化和低纹理场景,可以满足煤矿井下移动机器人的定位精度和稳定性要求。
