Multi-object tracking method based on deep reinforcement learning

Abstract

The invention discloses a multi-object tracking method and device based on deep reinforcement learning. The method comprises the following steps: extracting pedestrian characteristics; inputting the pedestrian characteristics to a prediction network to predicate pedestrian positions; and according to the pedestrian positions, obtaining pedestrian information, and inputting the pedestrian information to a decision network for judgment to track objects. The method can utilize information interaction between different objects and environment, and greatly improves tracking precision and performance.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a multi-target tracking method and device based on deep reinforcement learning. Background technique [0002] MOT (Multi-Object Tracking, multi-object tracking technology) has in-depth applications in various aspects such as video surveillance, human-computer interaction, and automatic driving. The purpose of multi-object tracking is to estimate the trajectory of different targets in the video and track them. Although there are many methods about MOT, which have been continuously proposed, it is very difficult to solve this problem in many unconstrained scenes, especially in crowded environments, because of the occlusion and huge intra-class differences between different objects. due to. [0003] In related technologies, multi-target tracking technologies can be mainly divided into two categories, the first one is offline type (also called batch processing type)...

AI Technical Summary

Problems solved by technology

[0004] However, the tracking accuracy of these methods is not very high, mainly because these methods are sensitive to occlusion and noise, such as missed detection, false detection and inaccurate labeling, etc.

Images