Pedestrian re-identification method and device, medium and mobile robot

Abstract

The invention discloses a pedestrian re-identification method and device, a medium and a mobile robot. The method comprises the following steps: constructing a pedestrian re-identification network model; training a backbone network and a detection head module of the pedestrian re-identification network model through the first training set; fixing network parameters of the optimized backbone network and the detection head module, and training an interested pooling module and a pedestrian feature extraction module of the pedestrian re-identification network model through a second training set; and collecting real-time pedestrian pictures, and detecting a target pedestrian based on the trained pedestrian re-identification network model. A pedestrian recognition network, a pedestrian feature extraction network and a pedestrian feature comparison module are integrated into a deep neural network, and operations are all transplanted to a graphics processor, so that an unnecessary data copying process is removed, the occupancy rate of CPU (Central Processing Unit) resources is reduced, the recognition speed of a target pedestrian is accelerated, the stable time consumption of the network is ensured, and the method and device easier to implement in an actual service environment.

Description

technical field [0001] The invention relates to the field of robots, in particular to a pedestrian re-identification method, device, medium and mobile robot. Background technique [0002] With the rapid development of the robot industry, various service robots emerge in an endless stream, and robots are more and more widely used in our life and work. In buildings in service industries such as hotels and banks, welcome robots can assist in the registration of guest identities, assist guests in storing their backpacks and other items, and follow guests until the end of the service. However, in actual scenarios, especially in scenarios where there is a large flow of people and people pass by each other, the robot is likely to misidentify the target guest or miss the target guest when the target guest is turning. The first prerequisite for the robot to follow guests is to re-identify pedestrians. Pedestrian re-identification is a technology to judge whether there is a specific ...

AI Technical Summary

Problems solved by technology

[0003] The invention provides a pedestrian re-identification method, device, medium and mobile robot, which solves the technical problems of low efficiency and long time delay in the existing pedestrian re-identification method

Images