Machine vision-based illumination robust identification and positioning method for object on plane

The invention relates to a machine vision-based illumination robust identification and positioning method for an object on a plane. The method comprises the following two parts: template generation and real-time detection; according to template generation, one of similar objects is selected as a standard model of the similar objects for template creation; according to real-time detection, other similar objects are identified and positioned through a pre-created template. By means of the method, objects in irregular shapes on the plane of a conveying belt can be identified and positioned, posesof the objects on the plane can be precisely calculated, the method can adapt to tasks of frequently replacing operation target types, the application range of robots in the field of industrial production is expanded, and the method has very important significance in improving the industrial production automation flexibility. 本发明涉及一种基于机器视觉的光照鲁棒的平面上物体识别定位方法，包括模板生成和实时检测两个部分：模板生成是由同类物体中选取其中一个作为该类物体的标准模型，进行模板创建；实时检测是通过事先创建好的