Research Of Liquid Leechdom Impurity Inspection Algorithm Based On Machine Vision

Liquid leechdom is applied in clinic extensively, and the leechdom quality is vital for patient. A fleetly and effective inspection of leechdom quality on-line is greatly increased along with the development of modern enterprise. Machine vision which gained the rapid development in recent years can provide effective technical means for production inspection. This dissertation emphasizes on the inspection algorithm of Liquid leechdom impurity, and the theory and method of the inspection system is expounded in detail.The basic concepts and the principle of machine vision, and at the same time, some algorithms about machine vision are discussed in this dissertation. An inspecting scheme based on machine vision is proposed with the use of industrial computer and high-speed camera, and some related contents about the system such as inspecting principle, system structure and function are discussed in the dissertation.The clear original image is the premise to inspect the target fast. With the foundation of the methods of analyzing and comparing the image smoothing, sharpening and histogram transform, the image enhancement algorithm based on histogram transform is proposed in allusion to the features of processing images. Through mapping gray levels, the algorithm enhances the contrast between the target and background and provides a good foundation for the further processing of the images.The image segmentation is the foundation to extract the inspecting targets correctly. Based on the analysis and research of the common segmentation algorithms, a adaptive threshold segmenting algorithm of the dynamic 2-D maximum entropy is proposed and used to segment the moving image which is with separated frame difference. The area and number of the inspection target is received according to the area mark of the segmentation result and furnish a judgment of the quality of products with a criterion. The segmentation algorithm based on optical flow is studied to contrast the segmentation results, and the result indicates that the optical flow method can effectively restrain light spot which is caused by the uneven wall of the bottle, but it should be studied further from the real-time aspect.In order to assess the availability of the research in this paper, Matlab is used to simulate. The result shows that this detection algorithm can satisfy the requirements of impurity inspection and achieve the expecting goals.