An Evolutionary Target Pose Measurement Based on Camera Motion Compensation

This paper centers on how to quickly and stably measure the target's position when the viewpoint is in motion. First, the target CAD model is constructed to obtain the edge template, the Chamfer distance setting fitness function is improved to evaluate the degree of matching between the edge template and the object, and the function is optimized and solved using genetic algorithm (GA). Secondly, the association model between the target pose and camera motion is established, and the image homography matrix and uniqueness constraints are used to solve the motion parameters of the camera, which is used to set the evolution strategy of the population in the GA and named as camera motion self-compensating genetic algorithm (CMSC-GA). The method is experimentally verified to improve the accuracy and speed of the target object position measurement.