基于毫米波雷达和CCD摄像机信息的D-S融合方法

D-S Fusion Method of Millimeter Wave Radar and CCD Camera

研究了毫米波雷达和CCD摄像机两种传感器信息的D-S融合方法,论述了毫米波雷达信息采集、滤波、计算障碍物的距离和方位及判别障碍物是否在无人地面平台ROI(Region of interested)内;CCD摄像机的信息处理、分割和障碍物的宽度、高度和方位角的计算。通过建立坐标系和坐标系间的齐次变换实现其空间配准。按D-S融合方法首先建立障碍物鉴别框架,通过实验确定它们对应于鉴别框架中障碍物指数的基本概率分配值,再根据D-S证据方法的组合规则计算所有证据联合作用下的融合结果,得到障碍物在无人地面平台ROI内的距离、方位和大小。在无人地面平台ROI内进行验证,本文的研究成果可行,并且具有较好的鲁棒性。

The sensor information D-S fusion method of millimeter wave radar and CCD camera is researched.This paper discusses information collection and segmentation of millimeter wave radar,computing the distance and orientation of the obstacle,judging whether the obstacle is in region of interested（ROI） of unmanned ground platform or not,information collection and segmentation of CCD camera,and calculation of width,height and azimuths of obstacle.Space matching is realized through establishing coordinate system and homogeneous transformation among coordinate systems.According to D-S fusion method,obstacle recognition frame is established at first,and its basic probability distribution value of obstacle index is ascertained by test.Then fusion results under evidence associative action are calculated according to combination rule of D-S evidence method,and distance,orientation and size of the obstacle in unmanned ground platform ROI are acquired.At last,the method is verified in unmanned ground platform ROI,and result shows that the method is validate and more robust.