Vision-based Detection Of Train Inverse Based On Image Dehazing Method

Dumper is a kind of large mechanical device that dumps material inside the train. Ifthere are train hook death and dumper continuing to work, it might damage the hook, notonly influence the production efficiency, but also caused huge economic loss. At present,people have designed a kind of detection system based on image processing method toautomatically identify the train hook death, the system use C80B model car which canidentify the death hook according to color of the train. Because the detection is based onimage processing, recognition accuracy could be easily influenced by light and weatherconditions, especially in fog weather, and fog weather occurs more and more frequently inChina’s As a consequence, besides conventional die hook test, haze removal should bestudied based on our application to ensure the stability of detection system in harshclimatic conditions.In this thesis, several image dehaze methods are presented and their advantages anddisadvantages were discussed firstly. Then, the train model and overall design of the traininverse detecting system were presented. After reference study, the research goal of thispaper is focused on image contrast of dark grey prior. Finally, image dehaze algorithm isimplemented in MATLAB2007b and the whole detection system is performed based onMATLAB platform and Visual C++platform.In our train inverse detection system, besides conventional train inverse detection,dark gray channel of image is measured to decide whether this image needs to be dehazed.If yes, image is reduild to remove fog based on estimation of transmission andatmospheric light. Furthermore, in order to meet the demand of real-time, the dark channelprior and oriented filtering algorithm to realize the combination of dehaze.The simulation results show that after combined with image dehaze method,vision-based train inverse detection system can accurately detect train inverse even underfog weather condition, esure the orderly of coal transportation in our country.