Cross-store operation behavior detection method based on target position information reasoning

Abstract

The invention discloses a cross-store operation behavior detection method based on target position information reasoning, and the method comprises the steps: employing a data set to train a target detection network, and obtaining a pedestrian and out-of-store article detection model and a storefront detection model; intercepting a monitoring video to obtain a frame image, and inputting the frame image into the pedestrian and out-of-store article detection model and the storefront detection model to obtain detection results corresponding to the two models; when pedestrians, storefronts and out-of-store articles appear in the detection result at the same time, the association relationship between the pedestrians and the out-of-store articles is judged according to the target position information, and the pedestrians with the association relationship and the out-of-store articles are considered as association targets; and for the associated target pedestrian, calculating a time weight ofthe associated target pedestrian, and when the time weight of the associated target pedestrian satisfies a cross-store operation behavior judgment condition, determining that the cross-store operationbehavior exists. The method can achieve the real-time and accurate detection of the cross-store operation behavior.

Description

technical field [0001] The invention belongs to the field of target recognition, and in particular relates to a method for detecting cross-store operation behavior based on target position information reasoning. Background technique [0002] Object detection has always been a research hotspot in the field of computer vision. Fast and accurate detection of desired targets is the direction of many scholars' efforts. A good recognition model requires fast and real-time detection of targets, and accurate positioning and classification of targets. In recent years, with the development of the neural network model and its excellent performance in the field of target detection, it has gained great attention in this field and a large number of application models have also been born. At present, there are still many object detection improvement model papers submitted and published at the annual computer vision conference. Mask R-CNN is two target detection models that perform well ...

AI Technical Summary

Problems solved by technology

[0006] At present, the target detection model has the characteristics of fast detection and is widely used. However, in the real scene, the cross-store business pedestrians outside the store are mobile and complicated. A single target detection model is not enough to meet the detection and judgment of cross-store business behavior. Therefore, there is an urgent need for a detection method that can quickly and accurately identify cross-store operations

Images