Method, device and equipment for detecting and tracking target objects in video

Abstract

The present application discloses a method, device and equipment for detecting and tracking a target object in a video, wherein the method specifically includes inputting continuous video frames into a pre-trained convolutional neural network; wherein the convolutional neural network is at least Including a set number of shared convolutional layers, a region proposal network layer, using the shared convolutional layer to extract the features of continuous video frames, and obtaining feature maps corresponding to different video frames respectively, using the region proposal network layer, according to The feature map determines a target area related to the target object, and based on the target area, the position and trajectory of the target object in consecutive video frames are detected. Through this method, the convolutional neural network model is used to unify the detection and tracking, reduce the amount of calculation, and solve the problem of multi-pose and multi-view detection of the target, improve the target detection rate, and reduce the false detection rate.

Description

technical field [0001] The present application relates to the technical field of machine vision, and in particular to a method, device and equipment for detecting and tracking a target object in a video. Background technique [0002] At present, the detection and recognition technology of dynamic target objects in video (such as: vehicles, pedestrians, etc. in video) is widely used in many fields such as transportation, security monitoring, and home entertainment. [0003] In the prior art, in order to extract a dynamic target object in a video, a background modeling and foreground detection (Background Generation And Foreground Detection) method is usually used to obtain a moving foreground, thereby obtaining a related target object. [0004] However, the above methods in the prior art are greatly affected by environmental factors such as illumination, which may cause more false detections, and may miss detections for stationary objects. Contents of the invention [0005...

AI Technical Summary

Problems solved by technology

[0004] However, the above-mentioned methods in the prior art are greatly affected by environmental factors such as illumination, which may cause more false detections, and may miss detection of stationary targets.

Images