摘要
针对采用当前算法进行机器人视觉定位时,存在定位朝向误差大和定位距离误差大的问题,提出多路径寻迹机器人主动视觉定位算法,由此提出多路径寻迹机器人主动视觉定位算法设计。通过增量式直方图计算算法优化粒子滤波的预测步骤和更新步骤,应用边缘方向直方图和颜色直方图进行计算粒子对应的权值,通过边缘方向直方图和颜色直方图实现目标跟踪。在机器人主动视觉定位过程中对摄像机进行标定,根据目标跟踪结果对视觉信息进行预处理,转换颜色空间,划分机器人识别区域,计算机器人的方向角和距离,实现机器人主动视觉定位。仿真结果表明,所提算法的定位朝向误差小、定位距离误差小。
In current algorithms,the orientation error and positioning distance error are too large.Therefore,an active vision positioning algorithm for multi-path tracking robot was put forward.The incremental histogram algo-rithm was used to optimize the prediction steps and update steps of particle filter.The edge orientation histograms and color histograms were used to calculate the corresponding weights of particles.On this basis,the target tracking was achieved.In the process of active vision positioning,the camera was calibrated.According to the tracking result,the vision information was preprocessed.Moreover,the color space was transformed,and the recognition regions were di-vided.Finally,the orientation angle and distance were calculated.Thus,the active vision positioning of robot was a-chieved.Simulation results show that the orientation error and distance error of the proposed algorithm are small.
作者
史梓潭
杨智勇
SHI Zi-tan;YANG Zhi-yong(College of Mechanics,Hubei University of Technology,Wuhan Hubei 430068,China)
出处
《计算机仿真》
北大核心
2020年第12期283-286,306,共5页
Computer Simulation
基金
国家自然科学基金项目(51907055)
湖北省教育厅中青年人才项目(Q20191405)。