Research On The Development Method About Robot Vision Based On Memory Mechanism

Based on the working principle of the structure of human cerebral cortex and information memory mechanism,the research of robot visual development method based on memory mechanism is studied.In the research,the paper realizes the online perception and learning of the scene by the human simulated intelligent image processing method,and optimizes the framework of object detection and tracking.The results of the research are as follows:1)For target tracking drift or lost in ambient illumination changes,posing variations and target occlusions,a multi feature weighted target tracking algorithm based on human memory mechanism is proposed.The experimental results show that the algorithm can effectively enhance the robustness of the features and show the learning and classification of memory database.2)A Hof forest target tracking method based on human memory mechanism is proposed.The online Hof forest decision tree is introduced to deal with the relationship between the image block and the target center,which is integrated into a variety of sample space position information and avoid over fitting;Three working memory spaces are introduced into the mechanism of human brain,which are instantaneous,short time and long time working memory.3)The algorithm is applied to the tracking target based on human memory mechanism.The tracking results verify the practicability of the proposed tracking algorithm,prove that the human memory based mechanism has broad application prospects.Robot visual development method based on memory mechanism imitates the human brain perception information in the aspect of image processing,optimizes the target detection and tracking framework,improves the robustness and effectiveness of traditional compressed sensing tracking algorithm.