Design And Realization Of Wrap Detect System Based On DSP' Machine Vision

With the development of computer, network and Image-processing, machine vision is playing a role in many aspects of the life. Such as physic image, photographic remote sensing, industrial-inspection and so on. So, as you know, the machine vision is almost needed in every place where the eyes may be used.As an application of on-line industrial-inspection, machine vision is researched in this thesis, hardware and software is designed, debugged and implemented with structured and modularized scheme. A small real-time wrap detect system which includes adjustable sampling rate and time of exposure is developed, and it overcomes many drawbacks of traditional wrap detect system so as to make the maintenance simple.The thesis analyses the reason of wrap, the merits and drawbacks of traditional wrap detect methods. According to the analysis, a new detect system based on smart camera VCSBC4018 is proposed. There are two parts in the system design. As to software, under the IDE CCS2.2, many modules such as image acquisition and storage, image-processing, wrap detect are designed. It also proposes a new method which both use crack width and the changing current of crack width to solve the problems caused by swinging tow. This method can reflect the the changing current of crack width which is better than using crack width alone. In order to adjust the sampling rate and time of exposure, a timing pulse whose period and width can be adjusted is developed, it is a trigger signal for the smart camera. With a group of emulation images, the result shows that the qualification of system meets the design requirement.In order to adjust the position of camera, a rectification software for camera is designed. It transfers the image from server to client with socket, then, the image will be displayed on the PC screen dynamically. As a result, the camera position can be put right. The software also can be used as monitored software.