Research Of Bar Code Image Real-time Processing System Based On DSP

This thesis deals with the application of image segmentation to the automatic identification of a bar code image with a complex background. The bar code image is segmented by the program which is coded by VC++ 6.0, then by image erosion, the background such as characters etc is processed, yielding isolated points which are filtered by a median filter. The relative position of the bar code does not change. Using template matching, the bar code region can be searched and segmented, then recognized by calling the recognition module.Because the DSP system has the very powerful function of computing and can fulfill the image processing in very short time, an experimental real-time processing system for bar code images is developed by using a DSP based chip-TMS320VE5410 in the thesis. The revolved hardware circuits, revolved softwares and design notices of the system are given.The bar code image recognition program coded by VC++6.0 is compared with the experimently real-time processing system, and the results show mat the processing speed of the latter is very greater than the speed of the front Studies in the thesis promote the automatic identification of bar code image and its practicalization and present some ideas of how to improve the bar code image processing system based on DSP. The mathematical models and algorithms for image erosion and image matching are given.