Pedestrian detection method based on layered kernel sparse representation

Abstract

The invention discloses a pedestrian detection method based on layered kernel sparse representation. The method is characterized by comprising steps as follows: preprocessing acquired traffic videos to obtain required positive and negative samples, acquiring multi-scale eigenvectors through layered sub-block division, and constructing two kinds of dictionary matrixes; preprocessing to-be-detected pedestrian images to obtain test samples, and performing pedestrian feature extraction with a method adopted in the dictionary construction process to form eigenvectors of the test samples; performing kernel sparse decomposition on the eigenvectors of the test samples with histogram cross kernel functions, adopting Gaussian functions for weighting in an iteration solution process, and reconstructing errors to realize pedestrian detection. With the adoption of the method, the higher detection performance can be realized, the pedestrian detection accuracy under the condition of partial shielding is effectively improved, and the robustness of a pedestrian detection system for appearance changes is enhanced.

Description

technical field [0001] The invention belongs to the field of intelligent transportation, and in particular relates to a pedestrian detection method based on hierarchical kernel sparse representation. Background technique [0002] Pedestrian detection is to separate the pedestrians appearing in the video or image from the background and precisely locate them. It has broad application prospects in video surveillance, intelligent driving and other fields. However, due to the large changes in the size, clothing, posture, or angle of view and illumination of pedestrian targets, coupled with complex background scenes and the movement and shaking of the camera itself, pedestrian detection requires high precision and real-time performance. , making pedestrian detection one of the most difficult topics in the field of intelligent transportation. [0003] At present, the pedestrian detection system is generally divided into two parts: appearance feature extraction and classification ...

AI Technical Summary

Problems solved by technology

[0004] In applications such as driving assistance systems, there are high requirements for the accuracy of pedestrian detection, but the existing methods generally use a fixed algorithm process to extract the global features of the target. For example, the implementation method of the direction gradient histogram is to first divide the image into small squares The unit connected area; then collect the gradient direction or edge direction histogram of each pixel in the grid unit; finally, these histograms can be combined to form a feature descriptor. The global features extracted by this inherent process cannot adapt to scenes and pedestrians. quick change of appearance

At the same time, due to the partial occlusion formed by street appendages, vehicles, trees, etc., it is also difficult for the classification model trained by the classifier to accurately distinguish pedestrians. It is difficult for existing methods to stably deal with complex situations such as partial occlusion and viewing angle change.

Images