The Technology Of Railway Track Surface Defect Inspection Based On DSP Image Processing

With the increasing of train velocity, there is a rising need for on railway safety, and therefore it is inevitable to have a better inspection for railway surface defect. This thesis tries to express building a comprehensive railway track surface inspection system basically through image processing and DSP. This system could prevent serious accident by real time monitoring the condition of railway tracks and reporting any surface defects. Thesis covers mostly cover following aspects:(1)Analyzes about several domestic and international railway surface defect inspection methods and discusses the major benefits of using image processing and DSP for railway inspection.(2)Designs an image pre-processing algorithm based on railway track image, finds a new method to take foreground out more efficiently, and tests this method in MATLAB.(3)Discusses about railway track surface defect recognition algorithm, classifies common surface defects and studies about different edge-detector and threshold-division methods. Chooses, based on railway track image, suitable edge-detector and threshold-division method. Runs simulation tests about surface defect recognition algorithm in MATLAB.(4)Generates a basic railway track surface defect inspection system frame work design, and completes hardware component selection according to the need.(5)Translates and implements all algorithms into C language under CCS3.3, designs a new UI and solves 24-bit BMP images loading problem. Generates a similar result but has a much better time efficiency compared with similar tests in Matlab, and therefore proves DSP has remarkable advantages on image processing of railway track surface inspection.