Pedestrian flow statistics method based on deep learning and multi-target tracking

Abstract

The invention relates to the technical field of image processing, and aims to provide a pedestrian flow statistics method based on deep learning and multi-target tracking. The method mainly comprisesthe following steps: S1, shooting a pedestrian monitoring video and reading an image in the video; S2, setting an effective area and a flow count of the image; S3, constructing a pedestrian detectionmodel based on deep learning and training the pedestrian detection model; S4, performing current pedestrian detection to obtain coordinates and image blocks of a current pedestrian frame; S5, trackingthe current pedestrian by using a multi-target tracking algorithm based on deep learning, and generating coordinates of the current pedestrian; S6, generating a moving track of the current pedestrian; S7, judging whether the current pedestrian leaves the effective area or not; If yes, entering the step S8, and if not, entering the step S4; S8, a noise threshold value is selected, and noise judgment is carried out; And S9, deleting the coordinates of the current pedestrian in the continuous video frames. According to the invention, an accurate flow statistics result can be provided in an actual use scene.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a pedestrian flow statistics method based on deep learning and multi-target tracking. Background technique [0002] The popularity of surveillance cameras and the development of image processing technology provide good conditions for intelligent surveillance. Pedestrian flow statistics are widely used in intelligent monitoring and can be used in various occasions such as hospitals, passages, and shops. An accurate traffic information can help to make reasonable resource allocation, estimate store rent levels and operating conditions, etc., which is of great significance. [0003] At present, there are many methods for people counting, one is to rely on hardware sensor equipment, and the other is to directly process the video. Among them, the sensor is greatly affected by the flow density of people; and the methods for counting the flow of people in video processing mai...

AI Technical Summary

Problems solved by technology

Among them, the sensor is greatly affected by the flow density of people; and the methods for counting the flow of people in video processing mainly include the following: a. The method based on face recognition, which is affected by face occlusion and head posture; b. Based on the head The method of detection and tracking, which places the camera on the top, has requirements for the installation environment, and at the same time, the occlusion of hats will have a greater impact, and the tracking algorithm is relatively simple and easy to cause tracking errors; c. Based on the method of head and shoulders detection and tracking, tracking The algorithm is relatively simple, but it is prone to errors due to the high density of pedestrians; d. The method based on human body matching and tracking, which uses clothing or some traditional features to judge pedestrians, which is prone to large errors; e. The method based on multiple cameras, the method Method requires multiple cameras for matching counting of pedestrians

In addition, the above-mentioned people flow counting method is basically based on the target moving trajectory crossing the auxiliary line to accumulate. However, due to the simple tracking algorithm adopted, it is easy to cause tracking confusion when there are many people, and the trajectory confusion will affect the counting, and the accuracy of pedestrian flow statistics is low.

Images