Automobile Connectors Recognition Algorithm Based on Improved Hu Invariant Moments

Aiming at the problems of low accuracy and slow speed of identification of automobile connectors produced in the assembly line, an improved Hu invariant moments algorithm for automobile connectors recognition is proposed. Firstly, the collected image of automobile connector is transformed into gray, filtered and de-noising, and threshold segmentation is used to extract the contour of automobile connectors, a gain function is designed to improve the recognition ability of Hu invariant moments for automotive connectors. Aiming at the problem that the traditional Hu invariant moments algorithm has a large amount of computation, the multi-scale feature expression of the image pyramid is used to reduce the image resolution in proportion and the amount of image traversal computation. At the same time, the problem of image feature loss caused by the high number of image pyramid layers is analyzed, an adaptive pyramid search strategy is proposed, the image distortion caused by the construction of pyramid is quantitatively analyzed, and the number of image pyramid layers is determined through the loss function. The experimental results show that compared with the traditional Hu invariant moments algorithm, the matching accuracy and speed of the proposed algorithm are improved by 5.03% and 45.06% respectively, and the recognition accuracy is greatly improved and the time is shorter, which provides an effective reference for automotive connectors recognition.