Research on the Method of Quickly Finding the Pedestrian Area of Interest

In pedestrian detection based on car video, it is necessary to quickly and accurately detect pedestrians. However, in the past the method that people used is exhaustive search. The method needs to detect all the areas in the picture, which is a waste of time. Based on the purpose of reducing the area to be tested, we use the edge detection and the principle of camera imaging, and run the test in MATLAB to get the method of finding the region of interest quickly. This method can reduce the retrieval area and shorten the retrieval time. The method can meet the requirements of real-time in pedestrian detection. Compared with the exhaustive search method, the number of windows that the method requires is 1/33 of the number of windows that the exhaustive search method requires. The detection speed of this method is several times higher than that of the exhaustive search method. It can be seen that the method is effective.