Research And Implementation Of The Intelligent Camera Oriented To Liquid Foreign Matter Detection

Drug solution plays a very important role in the health care. The quality of the drug solution not only affects the product quality,but also directly related to the effect of clinical treatment and the patient’s life safety.Detection of the foreign objects is an indispensable step in the test of pharmaceutical filling production line.So the research of intelligent camera based on the detection of the foreign objects in the drug solution has an important significance and value.This paper using the visible foreign objects in ampoules as study object.Firstly, we analyzes the key technology of visible foreign matter detecting system. According to the characteristics of the foreign objects itself, choose the appropriate way of light source and lighting to light the liquid;then study the algorithm which detect the foreign objects.After collected 3 frame image sequence through the image sensor, by comparing three kinds of moving object detection algorithm, combining background subtraction method and grayscale image morphological to obtain foreign objects movement area, as the histogram of foreign objects in the liquid showing unimodal, by simulation and analysis of several common image segmentation algorithms using matlab simulation tool, then proposed median otus adaptive threshold method based on particle swarm optimization algorithm and receive the relevant conclusions, in order to exclude the impact of liquid bubbles during detection of the visible foreign objects, according to the size of foreign objects and projection curve to identify foreign objects.Finally, design the intelligent camera detecting system based on the high-speed digital signal processor ADSP-BF533, the system includes video acquisition module,the CPU module,video display module and storage module.Through debugging foreign objects detection algorithm in the intelligent camera hardware platforms,realized the extraction of foreign objects.