Research On Pedestrian Detection And Tracking Algorithm Based On Multi-scale Method

Pedestrian detection is to extract pedestrians from static pictures or video sequences and accurately mark the specific location of pedestrians.Pedestrian tracking is to label pedestrians continuously in video sequence frames on the basis of accurate pedestrian detection.Based on the multi-scale method,this paper studies image preprocessing,feature extraction,classifier training,feature fusion and dimensionality reduction in the process of pedestrian detection and tracking.In the process of detection,the size of pedestrians is not clear in advance,and the multi-scale detection method can effectively realize the accurate detection of pedestrians at different scales in the image.In this paper,the multi scale sliding window method is used to extract the information of the images,scan,detect and identify the target pedestrians at different scales.Pedestrians have the characteristics of rigid and flexible,a single feature can’t fully describe the pedestrian characteristics,which brings difficulties to the detection and tracking.This paper proposes an idea of multi feature fusion,which combines different features and enriches the ability of feature description.HOG、LBP and Haar are fused respectively.It is found that the complementarity of the features plays an active role in improving the accuracy of pedestrian detection.In order to obtain a more comprehensive description of the characteristics,a high dimensional feature vector is obtained by integrating HOG、LBP and Haar.The main component analysis method is introduced to reduce the dimension of the high dimensional features.Less redundant information and less expressive information can reduce the computational complexity.The simulation shows that multi feature fusion can improve the accuracy and can optimize the accuracy of pedestrian detection to a certain extent.The method of dimensionality reduction is beneficial to improve the detection efficiency.It is positive for the realization of pedestrian detection and tracking.