Method, apparatus and system for human body tracking processing

Patent metadata

Field	Value
Publication number	US-11263445-B2
Application number	US-201916459868-A
Country	US
Kind code	B2
Filing date	Jul 2, 2019
Priority date	Jul 4, 2018
Publication date	Mar 1, 2022
Grant date	Mar 1, 2022

Abstract

A method, apparatus and a system for human body tracking processing, where an apparatus for video collection processing in the system has a built-in intelligent chip, and before uploading video data to a cloud server, the intelligent chip performs a pre-processing on the video data, retains a key image frame and performs a human body detection and a tracking processing on the key image frame by using human body detection tracking algorithm to acquire a first human body detection tracking result. Afterwards, the intelligent chip sends the first human body detection tracking result to the cloud server, so that the cloud server performs a human body re-identification algorithm processing and/or three-dimensional reconstruction algorithm processing on the first human body detection tracking result to acquire a second human body detection tracking result.

First claim

What is claimed is:

1. A method for human body tracking processing, wherein the method is applied to an apparatus for video collection processing which comprises a camera and an intelligent chip, and the method comprises:
collecting, by the camera, video data;
performing, by the intelligent chip, a human body detection tracking algorithm processing on the video data collected by the camera to acquire a first human body detection tracking result that comprises a human body foreground image, a two-dimensional position of a human body in a key image frame and a human body identifier; and
sending, by the intelligent chip, the first human body detection tracking result to a cloud server whereby the cloud server processes the first human body detection tracking result to acquire a second human body tracking result;
wherein the performing, by the intelligent chip, of the human body detection tracking algorithm processing on the video data collected by the camera to acquire the first human body detection tracking result comprises:
intercepting, by the intelligent chip, the video data collected by the camera to acquire a video image frame to be processed;
extracting the key image frame comprising the human body from the video image frame to be processed; and
performing a processing on the key image frame by using the human body detection tracking algorithm to acquire the first human body detection tracking result;
wherein the performing of the processing on the key image frame by using the human body detection tracking algorithm to acquire the first human body detection tracking result comprises:
performing a processing on the key image frame by using the human body detection tracking algorithm to acquire the human body foreground image, the two-dimensional position of the human body in the key image frame and the human body identifier: and
correspondingly, the sending, by the intelligent chip, the first human body detection tracking result to the cloud server, so that the cloud server processes the first human body detection tracking result to acquire the second human body tracking result comprises:
sending, by the intelligent chip, the human body foreground image, the two-dimensional position of the human body in the key image frame and the human body identifier to the cloud server, whereby the cloud server performs at least one of a human body re-identification algorithm processing and a three-dimensional reconstruction algorithm processing according to first human body detection tracking results sent by intelligent chips in at least two apparatuses for video collection processing to acquire the second human body detection tracking result.

2. A method for human body tracking processing, comprising:
receiving, by a cloud server, first human body detection tracking results sent by intelligent chips in at least two apparatuses for video collection processing respectively, wherein the first human body detection tracking results comprise human body foreground images, two-dimensional positions of human bodies in key image frames and human body identifiers, the first human body detection tracking results are obtained by performing a human body detection tracking algorithm processing on video data, and the video data is collected by cameras in the at least two apparatuses for video collection processing to which the intelligent chips belong; and
performing, by the cloud server, a processing on the first human body detection tracking results to acquire a second human body detection tracking result;
wherein the performing, by the cloud server, of the processing on the first human body detection tracking results to acquire the second human body detection tracking result comprises at least one of the following:
performing a re-identification on the human body foreground images in the first human body detection tracking results by using a human body re-identification algorithm to re-determine human body identifiers of the human bodies; and
performing, by using a three-dimensional reconstruction algorithm, a three-dimensional reconstruction on the two-dimensional positions of the human bodies in the key image frames in the first human body detection tracking results sent by the intelligent chips in the at least two apparatuses for video collection processing, wherein the human bodies are corresponding to a same human body identifier, and determine whether there is a human body identification error, and perform the re-identification on the human bodies when the human body identification error exists.

3. An apparatus for video collection processing, comprising:
a camera and an intelligent chip that are communicatively connected to each other;
wherein the camera is configured to collect video data; and
the intelligent chip is configured to perform a human body detection tracking algorithm processing on the video data collected by the camera to acquire a first human body detection tracking result that comprises a human body foreground image, a two-dimensional position of a human body in a key image frame and a human body identifier; and
send the first human body detection tracking result to a cloud server whereby the cloud server processes the first human body detection tracking result to acquire a second human body detection tracking result;
wherein the intelligent chip is configured to intercept the video data collected by the camera to acquire a video image frame to be processed, extract the key image frame comprising the human body from the video image frame to be processed, and
perform a processing on the key image frame by using the human body detection tracking algorithm to acquire the first human body detection tracking result;
wherein the intelligent chip is configured to perform a processing on the key image frame by using the human body detection tracking algorithm to acquire the human body foreground image, the two-dimensional position of the human body in the key image frame and the human body identifier; and
send the first human body detection tracking result to the cloud server whereby the cloud server performs at least one of a human body re-identification algorithm processing and a three-dimensional reconstruction algorithm processing according to first human body detection tracking results sent by intelligent chips in at least two apparatuses for video collection processing, so as to acquire the second human body detection tracking result.

4. A cloud server, comprising:
a processor and a memory as well as a communication interface connected to the processor;
the communication interface is connected to intelligent chips in at least two apparatuses for video collection processing;
the processor is configured to perform a method for human body tracking processing, comprising:
receiving, by the cloud server, first human body detection tracking results sent by the intelligent chips in the at least two apparatuses for video collection processing respectively, wherein the first human body detection tracking results comprise human body foreground images, two-dimensional positions of human bodies in key image frames and human body identifiers, the first human body detection tracking results are obtained by performing a human body detection tracking algorithm processing on video data, and the video data is collected by cameras in the at least two apparatuses for video collection processing to which the intelligent chips belong; and
performing, by the cloud server, a processing on the first human body detection tracking results to acquire a second human body detection tracking result; and
the memory is configured to store program code required by the processor for processing;
wherein the processor is configured to perform at least one of the following:
a re-identification on the human body foreground images in the first human b

Assignees

• Baidu online network technology beijing co ltd

Inventors

• Liu Zeyu
• Kang Le
• Zhang Chengyue
• Chi Zhizhen
• Wang Jian
• Li Xubin
• Liu Xiao
• Sun Hao
• Wen Shilei
• Ding Errui
• Zhang Hongwu
• Chen Mingyu
• Bao Yingze

Classifications

• G06V40/173

  face re-identification, e.g. recognising unknown faces across different face tracks · CPC title

• G06V40/103

  Static body considered as a whole, e.g. static pedestrian or occupant recognition · CPC title

• G06V20/52

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · CPC title

• G06T7/292 Primary

  Multi-camera tracking · CPC title

• G06F18/217

  Validation; Performance evaluation; Active pattern learning techniques · CPC title