CN106446958B - A kind of human body leaves reliable detection method - Google Patents
A kind of human body leaves reliable detection method Download PDFInfo
- Publication number
- CN106446958B CN106446958B CN201610881326.9A CN201610881326A CN106446958B CN 106446958 B CN106446958 B CN 106446958B CN 201610881326 A CN201610881326 A CN 201610881326A CN 106446958 B CN106446958 B CN 106446958B
- Authority
- CN
- China
- Prior art keywords
- human body
- human
- skin
- region
- colour
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/169—Holistic features and representations, i.e. based on the facial image taken as a whole
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Abstract
The present invention relates to a kind of human bodies to leave reliable detection method.Method includes the initialization of human testing region, based on the human body complexion extraction of colour of skin cluster, the human testing three parts based on color characteristic analysis.This method can reliably detect out human body target, while having faster processing speed using features of skin colors as human testing reference foundation.
Description
Technical field
The invention belongs to the field of video image processing towards intelligent monitoring, and in particular to a kind of human body leaves reliable detection
Method.
Background technique
Video monitoring is widely used in safety and supervision area, provides strong number for public safety and personnel's supervision
According to support.But the intelligent analysis degree at present towards monitor video is also relatively low, leaves in the human body towards monitor video
Context of detection there is no the disclosed technological means specifically for the application.Human body leaves detection and is mainly used for monitoring special scenes
Middle personnel play a significant role whether there is or not specified region is left in field of intelligent video surveillance.Human body towards monitor video leaves
Detection is the important application of monitor video image procossing.Its process flow are as follows: obtain image data from monitor video first, so
Detection zone is initialized afterwards and extracts human body complexion, further analyzes color characteristic in detection zone, and then realize that human body leaves
Detection.
The links in detection are left for human body, existing method such as patent 201510488408.2 uses frame difference method
Moving region in video is detected, can't detect if human body is stationary;Patent 201310405276.3,
201310116469.7 extracting characteristics of human body's structural classification device to be trained, patent 201110264004.7 combines background subtraction
With human testing classifier, patent 201010218630.8 is using the human body contour outline template detection multi-pose people with ambiguity
Body, patent 201310415544.X extract the feature obtained based on the colored human body detecting method with depth information, union feature
For human testing.Above method processing speed is very slow, is not used to the common computing platform such as DSP, ARM.Patent
201110026465.0 carrying out human testing based on depth image, it is not suitable for conventional monitor video image.
Summary of the invention
Detection problem is left for existing human body, on the basis of analysis of key link existing method is insufficient, the present invention is mentioned
A kind of human body towards intelligent monitoring leaves reliable detection method out, including human testing region is initialized, clustered based on the colour of skin
Human body complexion extract, based on color characteristic analysis human testing be total to three parts.The present invention is using features of skin colors as human body
Detection can reliably detect out human body target, while having faster processing speed referring to foundation.
Technical solution in the present invention is described below:
1, it is initialized based on the human testing region of Face datection and contouring head
A reference area must be specified by judging whether human body leaves, if using entire video pictures as detection zone,
The multiple human bodies of same picture then cannot be distinguished, and will increase calculation amount, reduce processing speed.In the present invention, it will examine for the first time
The upper half of human body region measured is as subsequent detection zone.As shown in Fig. 2, determining that specific step is as follows in the region:
Step1: using the preparatory trained Face datection classifier based on deep learning, in conjunction with contouring head template,
Detect human head location;
Step2: assuming that the head zone top left co-ordinate of human body is (x1,y1), wide and high respectively w1、h1;Human body it is upper
Half body region top left co-ordinate is (x2,y2), wide and high respectively w2、h2;Then according to data statistics rule, can substantially estimate
Upper half of human body position:
x2=x1-w1
y2=y1-h1/4
w2=w1×3
h2=h1×15/4
Innovative point is:
Directly extract that the traditional algorithm processing speed in human motion region is relatively slow and accuracy is not high, some calculations from video
Method is easy non-human moving target being determined as that human body, some algorithms are difficult to detect static human body.And according to Face datection knot
Syncephalon contouring can accurately detect static or movement human body target, accurately delimit out human body and leave detection zone
Domain.
2, the human body complexion based on colour of skin cluster extracts
The color of human body complexion and clothing, environment has larger discrimination, and the colour of skin of different people also slightly has difference.Based on this
Feature, can be according to human body skin tone testing human body target.After initializing human testing region, the colour of skin is used to the frame region
Cluster, obtains one group of characteristic value based on hsv color space, for describing human body complexion information.As shown in figure 3, specific steps
It is as follows:
Step1: color cluster is carried out to the human testing region after initialization;
Step2: skin color range is positioned according to head position, obtains Skin Color Information histogram feature;
Step3: being mapped to hsv color space for Skin Color Information histogram feature, obtains one group of concentration and describes Skin Color Information
Characteristic value.
Innovative point is:
It is clustered using the colour of skin, extracts the human body complexion feature based on hsv color model, accuracy in detection is high, processing speed
Fastly.
3, the human testing based on color characteristic analysis
Detection zone is left for human body, analyzes its color characteristic, is matched with human body complexion feature.As shown in figure 4,
Specific step is as follows:
Step1: it is directed to human testing region, is matched pixel-by-pixel with human body complexion feature;
Step2: statistics class colour of skin points then determine that human body leaves if fruit colour of skin points are less than threshold value Th;If detection zone
The width in domain, high respectively w2, h2, can value according to data statistics rule are as follows:
Th=(w2×h2)/45
Innovative point is:
Based on the human body complexion extracted come analysis detection region color feature, it can effectively detect human body target, mention
High human testing accuracy.
Detailed description of the invention
Fig. 1 is the overall schematic of the embodiment of the present invention;
Fig. 2 is the schematic diagram of human testing region initialization;
Fig. 3 is that human body complexion extracts schematic diagram;
Fig. 4 is the human testing schematic diagram based on color characteristic analysis.
Specific embodiment
Below with reference to diagram, the preferred embodiment of the present invention is described in detail.
It is as shown in Figure 1 that human body of the invention leaves detection workflow.A frame video image is read first;Then by pre-
First trained Face datection classifier combination contouring head template is handled, if detecting and navigating to head, basis
Its position initialization human body leaves detection zone;Then for the extracted region features of skin colors;Finally divide in every frame later
The region is analysed, verifies whether it meets color condition;If conditions are not met, then determining that human body leaves.
Human testing region initialization process is as shown in Figure 2.First using preparatory trained Face datection classifier, knot
Syncephalon contouring template, detects human head location;Then according to data statistic analysis, human body can be determined by head position
Leave detection zone.
It is as shown in Figure 3 that human body complexion extracts process.After initializing detection zone, color is carried out to detection zone first
Cluster;Then the range of colour of skin classification can be determined according to head position, and then seeks Skin Color Information histogram feature;Finally will
Histogram feature is mapped to hsv color space, obtains the characteristic value that one group of concentration describes Skin Color Information.
Human testing process based on color characteristic analysis is as shown in Figure 4.In every frame first after initialization, by human body
Detection zone and human body complexion feature carry out Similarity matching, judge whether it is class colour of skin point;Then the class skin of whole region is counted
Color dot obtains matching area, if it is less than threshold value, then determines that human body leaves.
It should be understood that for those of ordinary skills, it can be modified or changed according to the above description,
Such as change application field etc., and all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Technical solution in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is this hair
Bright a part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having
Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
Claims (2)
1. a kind of human body leaves reliable detection method, including the initialization of human testing region, the human body complexion based on colour of skin cluster
It extracts, the human testing three parts based on color characteristic analysis, which is characterized in that particular content is as follows:
(1), it is initialized based on the human testing region of Face datection and contouring head
A reference area must be specified by judging whether human body leaves, if using entire video pictures as detection zone, nothing
Method distinguishes the multiple human bodies of same picture, and will increase calculation amount, reduces processing speed, the upper half of human body that primary detection is arrived
Region determines that specific step is as follows in the region as subsequent detection zone:
Step1.1: using the preparatory trained Face datection classifier based on deep learning, in conjunction with contouring head template, inspection
Measure human head location;
Step1.2: assuming that the head zone top left co-ordinate of human body is, wide and high respectively、;The upper part of the body of human body
Region top left co-ordinate is, wide and high respectively、;Then according to data statistics rule, upper half of human body position is estimated
It sets:
(2), the human body complexion based on colour of skin cluster extracts
According to human body skin tone testing human body target, after initializing human testing region, the colour of skin is used to human testing region
Cluster, obtains one group of characteristic value based on hsv color space, for describing human body complexion information, the specific steps are as follows:
Step2.1: color cluster is carried out to the human testing region after initialization;
Step2.2: skin color range is positioned according to head position, obtains Skin Color Information histogram feature;
Step2.3: being mapped to hsv color space for Skin Color Information histogram feature, obtains one group of concentration and describes Skin Color Information
Characteristic value;
(3), the human testing based on color characteristic analysis
Detection zone is left for human body, analyzes its color characteristic, is matched with human body complexion feature, the specific steps are as follows:
Step3.1: it is directed to human testing region, is matched pixel-by-pixel with human body complexion feature;
Step3.2: statistics class colour of skin points, as fruit colour of skin points are less than threshold value, then determine that human body leaves.
2. a kind of human body according to claim 1 leaves reliable detection method, which is characterized in that the threshold valueValue
Are as follows: if width, the height of detection zone be respectively,, it is regular according to data statistics,Value are as follows:
。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610881326.9A CN106446958B (en) | 2016-10-09 | 2016-10-09 | A kind of human body leaves reliable detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610881326.9A CN106446958B (en) | 2016-10-09 | 2016-10-09 | A kind of human body leaves reliable detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106446958A CN106446958A (en) | 2017-02-22 |
CN106446958B true CN106446958B (en) | 2019-04-12 |
Family
ID=58172046
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610881326.9A Active CN106446958B (en) | 2016-10-09 | 2016-10-09 | A kind of human body leaves reliable detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106446958B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108416265A (en) * | 2018-01-30 | 2018-08-17 | 深圳大学 | A kind of method for detecting human face, device, equipment and storage medium |
CN111160169B (en) * | 2019-12-18 | 2024-03-15 | 中国平安人寿保险股份有限公司 | Face detection method, device, equipment and computer readable storage medium |
CN111653044A (en) * | 2020-04-26 | 2020-09-11 | 新石器慧通(北京)科技有限公司 | Automatic closing method and system for carrier accessories and unmanned vehicle |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030091472A (en) * | 2002-05-28 | 2003-12-03 | 엘지전자 주식회사 | Human face location detection method using adaptive color clustering |
CN103049747A (en) * | 2012-12-30 | 2013-04-17 | 信帧电子技术(北京)有限公司 | Method for re-identifying human body images by utilization skin color |
CN105469113A (en) * | 2015-11-19 | 2016-04-06 | 广州新节奏智能科技有限公司 | Human body bone point tracking method and system in two-dimensional video stream |
CN105913460A (en) * | 2015-11-26 | 2016-08-31 | 乐视致新电子科技(天津)有限公司 | Skin color detection method and device |
-
2016
- 2016-10-09 CN CN201610881326.9A patent/CN106446958B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030091472A (en) * | 2002-05-28 | 2003-12-03 | 엘지전자 주식회사 | Human face location detection method using adaptive color clustering |
CN103049747A (en) * | 2012-12-30 | 2013-04-17 | 信帧电子技术(北京)有限公司 | Method for re-identifying human body images by utilization skin color |
CN105469113A (en) * | 2015-11-19 | 2016-04-06 | 广州新节奏智能科技有限公司 | Human body bone point tracking method and system in two-dimensional video stream |
CN105913460A (en) * | 2015-11-26 | 2016-08-31 | 乐视致新电子科技(天津)有限公司 | Skin color detection method and device |
Non-Patent Citations (2)
Title |
---|
Face detection system using HSV color model and morphing;Vandana S. Bhata* and Jagadeesh D. Pujaria;《International Journal of Current Engineering and Technology》;20131231;全文 |
基于视频的人体检测与计数技术研究;王红梅;《中国优秀硕士学位论文全文数据库》;20120515;全文 |
Also Published As
Publication number | Publication date |
---|---|
CN106446958A (en) | 2017-02-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104063722B (en) | A kind of detection of fusion HOG human body targets and the safety cap recognition methods of SVM classifier | |
CN102521565B (en) | Garment identification method and system for low-resolution video | |
CN105160319B (en) | One kind realizing that pedestrian knows method for distinguishing again under monitor video | |
CN110378179B (en) | Subway ticket evasion behavior detection method and system based on infrared thermal imaging | |
CN102542289A (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
CN104992453A (en) | Target tracking method under complicated background based on extreme learning machine | |
CN104657712A (en) | Method for detecting masked person in monitoring video | |
CN103310194A (en) | Method for detecting head and shoulders of pedestrian in video based on overhead pixel gradient direction | |
CN108648211A (en) | A kind of small target detecting method, device, equipment and medium based on deep learning | |
CN106446958B (en) | A kind of human body leaves reliable detection method | |
CN110991315A (en) | Method for detecting wearing state of safety helmet in real time based on deep learning | |
CN106204594A (en) | A kind of direction detection method of dispersivity moving object based on video image | |
Chen et al. | Research on image fire detection based on support vector machine | |
TWI415032B (en) | Object tracking method | |
CN106909883A (en) | A kind of modularization hand region detection method and device based on ROS | |
CN106599880A (en) | Discrimination method of the same person facing examination without monitor | |
CN106548131A (en) | A kind of workmen's safety helmet real-time detection method based on pedestrian detection | |
CN102340620B (en) | Mahalanobis-distance-based video image background detection method | |
CN114913598A (en) | Smoking behavior identification method based on computer vision | |
Shah et al. | Multi-camera handoff for person re-identification | |
KR20190009006A (en) | Real time multi-object tracking device and method by using global motion | |
CN103077383B (en) | Based on the human motion identification method of the Divisional of spatio-temporal gradient feature | |
CN107103301A (en) | Video object space-time maximum stability identification color region matching process and system | |
Boonsim | Racing bib number localization on complex backgrounds | |
Hu et al. | Fast face detection based on skin color segmentation using single chrominance Cr |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20220128 Address after: 518048 23C, Huajing building, CaiTian South Road, Futian street, Futian District, Shenzhen, Guangdong Province Patentee after: SHENZHEN HUAJINDUN INFORMATION TECHNOLOGY Co.,Ltd. Address before: 410000 room 503, building 3, Yijing building, Debang new village, No. 588, Deya Road, Sifangping street, Fuqu District, Changsha City, Hunan Province Patentee before: HUNAN SUIFUYAN ELECTRONIC TECHNOLOGY CO.,LTD. |