Research on Abnormal Target Recognition of Full Information Mobile Monitoring Based on Machine Vision

When the color of moving object is close to the background, the accuracy of moving object recognition is affected. So the method of moving object recognition based on machine vision is designed. In order to reduce the distortion of image edge position, the moving object is calibrated and corrected by vision. In order to reduce the influence of noise to a controllable range, the full information mobile monitoring image is enhanced to preserve the image details. The edge features obtained from view and template are calculated by moment, and the similarity is obtained. Then the contour feature of moving monitoring target is extracted based on machine vision. Segmentation of the background region, according to the moving object trajectory center point information such as speed, direction and so on to determine whether the trajectory is abnormal events. The proposed method is tested on INRIA dataset and Vehicle Reld dataset, and the results show that the proposed method can improve the accuracy and recall rate and has good detection performance.