Target matching and positioning method based on edge point local characteristic values

The invention discloses a target matching and positioning method based on an edge point local characteristic values and the method comprises the following steps: an offline learning process: extracting the characteristic information of a template image, constructing a shape descriptor, and building an R-table; in the online process, performing voting on template information in the R-table by usingthe constructed actually measured shape descriptor, and when a voting score is greater than a threshold value, indicating matching succeeds; and target positioning: solving affine transformation parameters between the template contour and the actually measured image contour through iteration by using the Turkey weight and the least square adjustment principle, and carrying out affine transformation on the template contour to realize accurate positioning of the target. A good positioning result can be obtained under the conditions of translation, rotation and partial shielding of the target. 本发明公开了一种基于边缘点局部特征值的目标匹配与定位