Fast pedestrian detection method based on aggregated channel features

Abstract

The invention discloses a fast pedestrian detection method based on aggregated channel features, which comprises an early-stage position calibration process and a later-stage position screening process. During the early-stage position calibration process, multiple pieces of channel feature information in an input video or an image are aggregated, an image pyramid of the input image or the video in scale space is built, an image for each channel feature is calculated for each scale, features as pedestrian existing judgment basis are recognized, and the pedestrian position is initially extracted. During the later-stage position screening process, a convolutional neural network classifier is used for further screening each position calibrated in the early stage, pedestrians appearing in the image or the video are further detected, and a detection result is obtained. By adopting the technical scheme of the invention, in the case of a large training data amount, the classifier can automatically select features with a good recognition ability, the features can serve as pedestrian judgment basis, the method has good robustness, and the pedestrian detection precision is improved.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and relates to a pedestrian detection method, in particular to a fast pedestrian detection method based on aggregated channel features, which calibrates the accurate position of pedestrians existing in an input image or video. Background technique [0002] In the research of computer vision, object detection technology has a relatively long research history, and the problem of pedestrian detection exists as a sub-problem, which refers to the process of completing the calibration of the pedestrian position in the input image or video through a series of automated processing procedures. , where the position calibration is represented by a rectangular frame, and the coordinate information of the upper left corner and the lower right corner (or the length and width of the rectangular frame) of the rectangular frame needs to be given. Due to the posture of pedestrians in the input image, the i...

AI Technical Summary

Problems solved by technology

The sliding window algorithm using gradient histogram (Histogram of Gradient, HOG) feature plus support vector machine (Support Vector Machine, SVM) can better overcome the influence of illumination, so it is widely used, but it does not perform well when dealing with occlusion problems.

Felzenszwalb and Girshick et al. found through research that the local deformable part model (Deformable Part Model, DPM) can be used to describe the relationship between the various parts of the pedestrian to be detected and the whole, so that the occlusion problem can be effectively overcome, but the calculation process of this method is relatively complicated. , which runs slower

Images