The Research Of Moving Vehicles Detection Technology Based On Mobile Vision

In recent years, image processing and computer vision have become a very popular research. Moving object detection and tracking, as the core technology, is greatly concerned. It is used in various fields, such as intelligent monitoring system, intelligent transportation, intelligent human-computer interaction system, aerospace and industrial automation. With the rapid development of computer technology, image processing and analysis speed has been significantly improved. Thus, the development of image processing and computer vision technology is promoted.The purpose of this paper is to complete detection and tracking of moving vehicles in the sequence of images. In moving target detection section, we combine HOG(Histograms of Oriented Gradients) and SVM(Support Vector Machine) to recognize and detect moving vehicles:firstly, extract HOG features from sample library; secondly, we use SVM to train these features and get a sample model; then, calculate HOG features in every picture; finally, compare with the sample mode which we have got, and output the detection result. In moving target tracking section, we use Kalman filter, it can predict the positions of moving vehicles in the next frame with using the previous frame and the current frame image detection result. In this way, we can track moving vehicles. Since the computation time of each frame is too long, so we design a new rapid detection method based on feedback information. It can reduce the detection time from903ms to283ms and keep a high detection rate, it is about94.90%.