视觉同时定位与地图构建(Simultaneous localization and mapping,SLAM)过程中,动态物体引入的干扰信息会严重影响定位精度。通过剔除动态对象,修复空洞区域解决动态场景下的SLAM问题。采用Mask-RCNN获取语义信息,结合对极几何方法对动...视觉同时定位与地图构建(Simultaneous localization and mapping,SLAM)过程中,动态物体引入的干扰信息会严重影响定位精度。通过剔除动态对象,修复空洞区域解决动态场景下的SLAM问题。采用Mask-RCNN获取语义信息,结合对极几何方法对动态对象进行剔除。使用关键帧像素加权映射的方式对RGB和深度图空洞区域进行逐像素恢复。依据深度图相邻像素相关性使用区域生长算法完善深度信息。在TUM数据集上的实验结果表明,位姿估计精度较ORB-SLAM2平均提高85.26%,较DynaSLAM提高28.54%,在实际场景中进行测试依旧表现良好。展开更多
A hybrid remotely operated underwater vehicle( HROV) capable of working to the full ocean depth has been developed. In order for the vehicle to achieve a certain survivability level,a self-repairing control system( SR...A hybrid remotely operated underwater vehicle( HROV) capable of working to the full ocean depth has been developed. In order for the vehicle to achieve a certain survivability level,a self-repairing control system( SRCS) has been designed. It consists of two basic technologies,fault diagnosis and isolation( FDI) and reconfigurable control. For FDI,a model-based hierarchical fault diagnosis system is designed for the HROV. Then,control strategies which reconfigure the control system at intervals according to information from the FDI system are presented. Combining the two technologies,it can obtain the fundamental frame of SRCS for the HROV. Considering the hazardous underwater environment at the limiting depth and the hybrid operating modes,an assessment of the HROV's survivability is vitally needed before it enters operational service. This paper presents a new definition of survivability for underwater vehicles and develops a simple survivability model for the SRCS. As a result of survivability assessment for the SRCS,we are able to figure out the survivability of SRCS and make further optimization about it. The methodology developed herein is also applicable to other types of underwater vehicles.展开更多
文摘视觉同时定位与地图构建(Simultaneous localization and mapping,SLAM)过程中,动态物体引入的干扰信息会严重影响定位精度。通过剔除动态对象,修复空洞区域解决动态场景下的SLAM问题。采用Mask-RCNN获取语义信息,结合对极几何方法对动态对象进行剔除。使用关键帧像素加权映射的方式对RGB和深度图空洞区域进行逐像素恢复。依据深度图相邻像素相关性使用区域生长算法完善深度信息。在TUM数据集上的实验结果表明,位姿估计精度较ORB-SLAM2平均提高85.26%,较DynaSLAM提高28.54%,在实际场景中进行测试依旧表现良好。
基金Sponsored by the National Natural Science Foundation of China(Grant No.51109132)Specialized Research Fund for the Doctoral Program of Higher Education of China(Grant No.20110073120015)
文摘A hybrid remotely operated underwater vehicle( HROV) capable of working to the full ocean depth has been developed. In order for the vehicle to achieve a certain survivability level,a self-repairing control system( SRCS) has been designed. It consists of two basic technologies,fault diagnosis and isolation( FDI) and reconfigurable control. For FDI,a model-based hierarchical fault diagnosis system is designed for the HROV. Then,control strategies which reconfigure the control system at intervals according to information from the FDI system are presented. Combining the two technologies,it can obtain the fundamental frame of SRCS for the HROV. Considering the hazardous underwater environment at the limiting depth and the hybrid operating modes,an assessment of the HROV's survivability is vitally needed before it enters operational service. This paper presents a new definition of survivability for underwater vehicles and develops a simple survivability model for the SRCS. As a result of survivability assessment for the SRCS,we are able to figure out the survivability of SRCS and make further optimization about it. The methodology developed herein is also applicable to other types of underwater vehicles.