Micro Parts Recognition And Tracking While Being Assembled

Micro assembly now is not only a hotspot of educational circles, but also an important technology which is able to bring many countries new economic growth points. The micro-nano products assembled using the micro assembly technology are set a high value. And the micro parts recognition and tracking technology is concerned because of its significant effect during assembling. Nowadays nearly all advanced micro assembly systems contain the micro-vision component to accomplish the supervision and measurement task instead of artificiality.After careful study of the characteristics of the whole micro assembly system and the image of the micro parts, this dissertation designed the multitarget recognition and tracking or locating algorithm based on the quality of inter-frame difference according to some traditional algorithms such as Kalman Filter, Inter-frame Difference and so on. This algorithm combined with Hu invariant moment is able to obtain the real time position of the taget, using the one-dimensional search and improved Powell algorithm to optimize the parameters of the rectangle tracking or locating frame. The data acquired by this algorithm is to be used for guiding the precision instruments and machineries to accomplish the assembly task. Experiments show that the algorithm designed by this dissertation is able to track and recognize or locate multitargets. The program for this algorithm is convenient to be reused and expanded because of using the OOD character of C++.Besides, this dissertation provided a detailed proposal for the usage of the software while assembling.