Face and pedestrian perception system based on depth camera

Abstract

The invention discloses a face and pedestrian perception system based on a depth camera. The system is applied to scenes including a mobile robot, intelligent monitoring and the like. The system at least comprises three modules, wherein a hardware layer module is used for accessing various heterogeneous cameras; a driving layer module is used for unifying various camera driving interfaces through an ROS (Read-Only-Storage) system; and an application layer module comprises a face / head detector based on a Dlib trainer and RGB(Red, Green and Blue)-D information, a pedestrian detector based on a multi-modal deep learning model and a pedestrian tracker based on a multimachine system. The system can be divided into a single-depth camera system and a multi-depth camera system by aiming at the number of cameras, the perception application of the single-depth camera system comprises face detection and identification and pedestrian detection and tracking; and the multi-depth camera system overcomes the defects of small coverage, a camera network is formed, and the cross-region long-time tracking of the pedestrian is realized.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a face and pedestrian perception system applied to a video surveillance scene or deployed on a mobile robot. Background technique [0002] Object detection is the main research basis and hotspot of computer vision system. As a basic task of computer vision, object detection is a research hotspot in academia and industry. Among them, human perception has a wide range of application significance, especially in industries such as intelligent security, unmanned driving and mobile robots. Many of the solutions in these industries use depth cameras for human perception to achieve fast, accurate and three-dimensional positioning. In some commercial and open source codes, it is generally only for a single device or a specific scenario, and it is difficult to carry out secondary development or function expansion and deletion according to user needs. Based on the above reasons...

AI Technical Summary

Problems solved by technology

[0004] Commercial products do not allow developers to develop their source code, which cannot satisfy developers to develop applications according to their own needs

There are also many open source object detection source codes on the Internet, but these source codes are basically only for a certain task or a single camera, and they are not modularized, so it is difficult to reuse them

Images