Algorithm research on detection region tracking of bottle mouth image in intelligent empty bottle inspection system

In the design of intelligent empty bottle inspection system, the effective means to improve the detection speed is processing only the region of interesting in the obtained image, however, the image has a certain shift because of mechanical vibration and so on. So improving the detection accuracy and speed is based on the premise that the Detection area is traced correctly. In this paper, a new tracking algorithm for detection region of bottle mouth image was proposed. First, circular scanning the linear domain, and classifying the marginal point twice, then fitting the marginal point classified, finally the position of bottle mouth can obtained. The experiments show that the tracking algorithm can improve speed and precision of positioning about bottle mouth image detection region, and it has great practical value.