Research of Automatic Detection System of Illegal Parting in Highway
|School||Henan University of Science and Technology|
|Keywords||Illegal parking Frog-degraded images Image enhancement Retinex algorithm Vehicle speed estimation|
Because of the characteristics of heavy in traffic volume and fast in speed of thehighway, the illegal parking behavior has been one of the major causes of serioustraffic accidents on highway. The illegal parking behavior detection system is animportant part of the highway video detection system, as well as the foundation of thehighway intelligent transportation system.Existing highway video detection system is able to detect the traffic volume andcongestion correctly and effectively in normal weather conditions. As a kind ofcommon natural phenomenon, the fog makes the atmosphere visibility reduced, andleads to the degradation in contrast and color of the outdoor scene. So the degradationof images captured by the cameras in highway is seriously, and the video detectionsystem is difficult to detect and track the moving vehicles in correct. The imageenhancement algorithms based on Retinex theory has been a hot area of research fordefogging. Variable filter Retinex algorithm is more effective among the imageenhancement algorithms based on Retinex theory. It is better in enhancing details andpreserving color fidelity. But the image enhanced by variable filter Retinex algorithmis overall dark and not high in contrast. This paper used variable filter Retinexalgorithm, with POSHE algorithm enhancing the contrast of the enhanced images, toenhance frog-degraded images.The speed of vehicles is one of the important parameters for judging the illegalparking. This paper mainly researched two effective video speed measuring method:measurement of vehicle speed based on Harris corner detector and measurement ofvehicle speed based on the estimation of feature optical flow, and analyzed theadvantages and disadvantages of the two methods. The speed of vehicle has beenmeasured by the later in engineering.