Research On The System For The Automatic On-line Visual Inspection Of Electret Condenser Microphone

Automatic visual inspection is an important research area in information processing field. The dissertation is based on the project of the System for Electret Condenser Microphone Performance Test, and adopts automatic visual inspection technology to solve the problem of micro volume accessory's angle detection and polar recognition. In the process of microphone manufacture, some inspections must be executed to insure the quality of the accessories. Because of the small volume of the accessory and high precision and rapid velocity request of the project, the microphone manufacture is crying for on-line visual inspection. The dissertation researched the algorithm of software implement and the method of hardware implement.The main work of the dissertation are summarized as follows:Firstly, the structure of the inspection system was introduced, especially the principle of image collection. All the work were based on the survey and study in some machine vision companies.Secondly, in the research of the. inspection algorithm we selected the appropriate method of image pretreatment algorithm which included the binarization algorithm and algorithms of extracting the object region and eliminating image noise and thorns. The thinning algorithm based on mathematical morphology and Hough transform algorithm were used to detect the angle of the line. The method of locating image centroid was used to recognize the character of the image.Lastly, the dissertation introduced the real-time image processing system. Hough transform algorithm was realized on DSP, and the result had checked the high efficiency and feasibility of DSP. It was a new method of improving efficiency of the inspection system.