Method, apparatus and system for acquiring headcount

• US 20150317517A1
• Filed: 09/03/2013
• Published: 11/05/2015
• Est. Priority Date: 11/26/2012
• Status: Active Grant

• Current Assignee: Xi'An Zhongxing New Software Co., Ltd. (ZTE Corporation)
• Original Assignee: ZTE Corporation
• Inventors: Wang, Dong, Zeng, Gang, Quan, Jiacheng
• Granted Patent: US 9,824,460 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/784 the detected or recognised objects being people
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 2207/30242 Counting objects in image
• CPC Class Codes: G06T 7/251 involving models
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06V 20/53 Recognition of crowd images, e.g. recognition of crowd congestion
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16
• CPC Class Codes: G07C 9/00 Individual registration on entry or exit
• CPC Class Codes: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

A method, apparatus and system for acquiring headcount are provided. Data frames are acquired from a surveillance video, object models are set in the data frames, and a foreground image is extracted from the data frames, wherein the foreground image includes moving objects extracted from the data frames; the moving objects in the foreground image is matched with the object models, the headcount in corresponding data frames is acquired according to a result of the matching, characteristics of the matched moving objects are extracted, and the corresponding moving objects are tracked based on the characteristics; the actual headcount for a predetermined time segment of the surveillance video is acquired based on a result of the tracking and the acquired headcount for each of the data frames, wherein the result of the tracking includes: an indication of whether some or all of the data frames includes a same moving object.

Claims

1. A method for acquiring headcount, comprising:

acquiring data frames from a surveillance video, setting object models in the data frames, and extracting a foreground image from the data frames, wherein the foreground image comprises moving objects extracted from the data frames;

matching the moving objects in the foreground image with the object models, acquiring the headcount in corresponding data frames according to a result of the matching, extracting characteristics of the matched moving objects, and tracking the corresponding moving objects based on the characteristics; and

acquiring the actual headcount for a predetermined time segment of the surveillance video based on a result of the tracking and the acquired headcount for each of the data frames, wherein the result of the tracking comprises;

an indication of whether some or all of the data frames includes a same moving object or not.

2. The method according to claim 1, wherein the method further comprises:

after extracting the foreground image from the data frames, eliminating interference in the foreground image.

3. The method according to claim 2, wherein the eliminating interference in the foreground image comprises:

eliminating shadow and/or noise in the foreground image by means of a difference in colour space of a Hue Saturation Value (HSV) between a current data frame image and the background image.

4. The method according to claim 1, wherein the method further comprises:

setting a detection line in the data frames after acquiring data frames from the surveillance video, wherein the result of the tracking further comprises;

a moving track of the same moving object in some or all of the data frames;

accordingly, wherein the method further comprises;

after acquiring the actual headcount for the predetermined time segment of the surveillance video, and

acquiring the actual headcount in different directions for the predetermined time segment of the surveillance video based on the result of the tracking and position of the tracked moving objects relative to the detection line.

5. The method according to claim 1, wherein the extracting a foreground image from the data frames comprises:

extracting the foreground image from the data frames by background modelling or by determining an inter-frame difference.

6. The method according to claim 1, wherein the characteristics of the matched moving objects comprise one of the following:

centre of the moving object, object window of the moving object, area of the moving object, grey-level histogram of the moving object, colourful histogram of the moving object, angular points of the moving object and angular points of the moving object at a sub-pixel level.

7. The method according to claim 1, wherein the acquiring the actual headcount for a predetermined time segment of the surveillance video comprises:

counting the actual headcount for the predetermined time segment of the surveillance video with a counter.

8. An apparatus for acquiring headcount, comprising:

an acquisition module, configured to acquire data frames from a surveillance video;

a setting module, configured to set object models in the data frames;

a foreground image extracting module, configured to extract a foreground image from the data frames, wherein the foreground image comprises moving objects extracted from the data frames;

an object matching module, configured to match the moving objects in the foreground image with the object models and to acquire the headcount in corresponding data frames according to a result of the matching;

a characteristic extracting module, configured to extract characteristics of the matched moving objects and to track the corresponding moving objects based on the characteristics;

a counting module, configured to acquire the actual headcount for a predetermined time segment of the surveillance video based on a result of the tracking and the acquired headcount for each of the data frames, wherein the result of the tracking comprises;

an indication of whether some or all of the data frames includes a same moving object or not.

9. The apparatus according to claim 8, wherein the foreground image extracting module is further configured to, after extracting the foreground image from the data frames, eliminate interference in the foreground image.

10. The apparatus according to claim 9, wherein the foreground image extracting module is configured to eliminate shadow and/or noise in the foreground image by means of a difference in colour space of a Hue Saturation Value (HSV) between a current data frame image and the background image.

11. The apparatus according to claim 8,

wherein the setting module is further configured to set a detection line in the data frames;

and

wherein the counting module is further configured to acquire the actual headcount in different directions for the predetermined time segment of the surveillance video based on the result of the tracking and position of the tracked moving objects relative to the detection line.

12. The apparatus according to claim 8, wherein the foreground image extracting module is configured to extract the foreground image from the data frames by background modelling or by determining an inter-frame difference.

13. The apparatus according to claim 8, wherein the characteristics of the matched moving objects comprise one of the following:

centre of the moving object, object window of the moving object, area of the moving object, grey-level histogram of the moving object, colourful histogram of the moving object, angular points of the moving object and angular points of the moving object at a sub-pixel level.

14. The apparatus according to claim 8, wherein the counting module is configured to count the actual headcount for the predetermined time segment of the surveillance video with a counter.

15. A system for acquiring headcount, comprising an apparatus for acquiring headcount and a data collector,

wherein the data collector is configured to collect data frames from a surveillance video, and

wherein the apparatus for acquiring headcount comprises:

an acquisition module, configured to acquire data frames from a surveillance video;

a setting module, configured to set object models in the data frames;

a foreground image extracting module, configured to extract a foreground image from the data frames, wherein the foreground image comprises moving objects extracted from the data frames;

an object matching module, configured to match the moving objects in the foreground image with the object models and to acquire the headcount in corresponding data frames according to a result of the matching;

a characteristic extracting module, configured to extract characteristics of the matched moving objects and to track the corresponding moving objects based on the characteristics;

a counting module, configured to acquire the actual headcount for a predetermined time segment of the surveillance video based on a result of the tracking and the acquired headcount for each of the data frames, wherein the result of the tracking comprises;

an indication of whether some or all of the data frames includes a same moving object or not.

16. The system according to claim 15, wherein the system further comprises:

a platform management server for intelligent analysis, one or more unit management servers for intelligent analysis and one or more algorithm analysis capability servers for intelligent analysis platform, wherein the apparatus for acquiring headcount is set in the one or more algorithm analysis capability servers for intelligent analysis platform;

wherein the data collector is configured to collect the data frames from the surveillance video and to send the collected data frames to the platform management server for intelligent analysis;

wherein the platform management server for intelligent analysis is configured to select a unit management server for intelligent analysis according to a headcount counting instruction from a client, to issue a first headcount counting task to the selected unit management server for intelligent analysis, and to report a result of the counting to the client, wherein the first headcount counting task carries the collected data frames;

wherein the unit management servers for intelligent analysis are configured to select a algorithm analysis capability server for intelligent analysis platform according to the received first headcount counting task, to issue a second headcount counting task to the selected algorithm analysis capability server for intelligent analysis platform, and to report a result of the counting to the platform management server for intelligent analysis, wherein the second headcount counting task carries the collected data frames; and

wherein the algorithm analysis capability servers for intelligent analysis platform are configured to enable the apparatus for acquiring headcount according to the received second headcount counting task so as to perform a headcount counting, and to report a result of the counting to the unit management server for intelligent analysis.

17. The system according to claim 15, wherein the foreground image extracting module is further configured to, after extracting the foreground image from the data frames, eliminate interference in the foreground image.

18. The system according to claim 17, wherein the foreground image extracting module is configured to eliminate shadow and/or noise in the foreground image by means of a difference in colour space of a Hue Saturation Value (HSV) between a current data frame image and the background image.

19. The system according to claim 15, wherein the setting module is further configured to set a detection line in the data frames;

and

wherein the counting module is further configured to acquire the actual headcount in different directions for the predetermined time segment of the surveillance video based on the result of the tracking and position of the tracked moving objects relative to the detection line.

20. The system according to claim 15, wherein the foreground image extracting module is configured to extract the foreground image from the data frames by background modelling or by determining an inter-frame difference.