A pedestrian detection method and system in a densely populated area

Abstract

The invention relates to a pedestrian detection method and system in a densely populated area. By improving the constructed YOLOv4 mark detection network, an attention mechanism module is added to increase the weight of pedestrian features, and a guide anchor frame module is added to adaptively adjust the anchor. The shape of the frame improves the accuracy of pedestrian detection while ensuring the speed of network reasoning, so as to solve the problem of application difficulties caused by the imbalance between the speed and accuracy of detection methods in the prior art. The present invention is improved on the basis of the YOLOv4 network, which not only retains the advantages of the YOLOv4 network itself, which is fast and easy to learn the generalization characteristics of the target, but also reduces the interference of invalid feature information by introducing the attention mechanism module, and effectively improves the performance of complex environments. The accuracy of pedestrian recognition; at the same time, the present invention adaptively adjusts the size of the anchor frame based on the guided anchor frame method, thereby generating an anchor frame more in line with the size of the human body, which can effectively improve the accuracy of pedestrian recognition.

Description

Technical field [0001] The present invention relates to the field of target detection, and more particularly to a human detection method and system for human flow intensive region. Background technique [0002] In recent years, the continuous development of economic and social development and urbanization have made more and more people leave their hometown and flour into the big city. The high gathering of the crowd has enabled the demand for the detection of dense regional pedestrians in public places. For urban rail traffic scheduling, real-time detection of pedestrians in the station can master the passenger flow density in the station, better deploy the vehicle, alleviate the peak passenger pressure; for the security of the subway station, accurate and quickly detect the pedestrians in the intensive area can be effective Understand the aggregation of the population, facilitate security scheduling, prevent safety accidents such as stepping. [0003] The method currently used f...

AI Technical Summary

Problems solved by technology

[0006] Based on the above situation of the prior art, the object of the present invention is to provide a method and system for pedestrian detection in densely populated areas, by improving the built faster YOLOv4 mark detection network and using it for pedestrian detection. Solve the problem of difficulty in practical application caused by the unbalanced speed and accuracy of detection methods in the prior art

Images