Moving Objects Detection And Tracking Of Video Images In Industrial Sewing Machines

With the development of the digital video, the detection and tracking of moving objects in video images is a study hotspot. Due to the diversity and complexity, there is not an all-purpose algorithm, so it is necessary to select or design suitable algorithms for specific applications.Moving object tracking of video images in industrial sewing machine is researched in this paper. The correlative algorithms of the moving object detection and tracking are applied to the system of sewing machine fault detection. The function that the moving of sewing machine'needle and many feature points are tracked is implemented.Depending on whether sewing machines'needle is blocked, different algorithms are applied to tracking of moving sewing machines'needle:When the sewing machine's needle is blocked, the template matching algorithm based on Hausdorff distance is improved in this paper, and it is applied to the tracking algorithm based on object shape in order to get the position of sewing machine's needle; When it not, a algorithm based on connected components is designed in order to track the moving sewing machine's needle.The algorithm based on image feature points is applied to tracking the moving of sewing machine's parts:firstly, Harris corners algorithm is applied to distill object's feature points; secondly, a matching and tracking algorithm of feature points suitable to the subject is proposed in this paper, and applied it and the algorithm of feature point distilling together in order to completing the sewing machine parts'tracking.The experiments showed that the moving of sewing machine's needle and parts can be tracked with high precision with the above methods so as to provide reliable and comprehensive data for analyzing for sewing machine fault diagnosis.