Detection Algorithm Of Illegal Occupation Of Bus Lanes Based On Vehicle Vedio

With the rapid development of China's economy, public transportation has also gone through great improvement, which provides great convenience for people's life. However, the phenomenon that bus lanes are occupied by private vehicles is becoming more serious with the increasing number of private vehicles. It is difficult for traditional fixed-point monitoring and manual monitoring to handle the increasingly serious phenomenon of illegal occupation of bus lanes. Therefore, intelligent transportation is becoming the vital way to deal with the traffic problems. The detection system of illegal occupation of bus lanes is an important part of intelligent transportation. The detection system is a new intelligent monitoring technology and it can recognize the bus lanes and vehicles breaking traffic rules automatically with the surveillance cameras in the bus. Compared with the traditional monitoring method, the detection system can reduce manual intervention and improve detections efficiency greatly.This paper introduces the composition and realization of the detection in detail. We use the improved Hough transform algorithm in lanes detection and recognition module. Once we recognize the lanes, we can take advantage of the difference between the colors of the normal lanes and bus lanes to gain SVM classifier and distinguish bus lanes in detection areas, and use a state machine to maintain the stability of detection. In license plate detection and identification module, we first use Adaboost cascaded classifier based on Haar-like feature to identify target license plates in detection areas. In case of detection errors, we further analyze the features of the edge and color of target license plates. Then we come up with Sobel edge feature verification and RGB feature verification to identify target license plates. Once license plates are identified as targets, we extract H(hue) feature of the license plate by transforming the license plate image into HSV color space. Then we put license plates into different categories according tocolors by using SVM classifier to judge whether target license plates are normal license plates or bus license plates. The LK optical flow method is used to track the detection results. We then continue to distinguish the location of license plates and bus lanes to identify the vehicles breaking traffic rules. The detection system which detects the illegal occupation of bus lanes can improve the supervision department's working efficiency greatly and reduce manual intervention. What's more, the detection system can guarantee the detection efficiency and accuracy.