A Hybrid Method to Detect and Verify Vehicle Crash with Haar-Like Features and SVM Over the Web

In this paper, a multi-level vehicle crash detection is proposed. The first step is to detect and track the movement of the vehicles in a scene and based on the stop condition of any vehicle, we trigger the first level which is a possible crash detected. Next we scan the object with a pre-trained cascade detector looking for possible defects in it. The areas detected are passed into a verification function which is based on a pre-trained SVM to classify their HOG if they represent area of damage in a vehicle or not. We tested static images of few cars and we were able to detect and verify the defect in their body. Then we built a webcam simulator and a web application to test the solution in a real-time environment and over the web. The test succeeded and we were able to detect a real accident successfully.