CN105631852B - Indoor human body detection method based on depth image contour - Google Patents
Indoor human body detection method based on depth image contour Download PDFInfo
- Publication number
- CN105631852B CN105631852B CN201510740255.6A CN201510740255A CN105631852B CN 105631852 B CN105631852 B CN 105631852B CN 201510740255 A CN201510740255 A CN 201510740255A CN 105631852 B CN105631852 B CN 105631852B
- Authority
- CN
- China
- Prior art keywords
- contour
- human body
- depth
- mrow
- image
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
- G06T2207/10012—Stereo images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a kind of indoor human body detection method based on depth image contour, its method and step are as follows：A, depth information image is obtained；B, depth image three-dimensional contour is obtained；C, top layer three-dimensional contour number of people centralized positioning；D, depth secondary detection；E, secondary high-rise human body centralized positioning；F, the contour line profile marked and built-in human body head depth image are subjected to silhouette contrast, exclude flase drop；H, by the calculating of above step C tetra- steps of F, the position of human body of different height in depth information image can be marked.Present invention, avoiding due to Consumer's Experience caused by detection method it is not good enough, it is poor for applicability, improve human testing speed, enhancing detection accuracy.
Description
Technical field
The present invention relates to pattern-recognition, machine learning field, more particularly to a kind of interior based on depth image contour
Human body detecting method.
Background technology
Indoor environment is people's work, study, an important place of rest.The human testing and counting of indoor environment,
It is the important component of intelligent video monitoring system, allows smart machine easily to obtain the human body quantity in indoor environment with dividing
Cloth situation, can make human body target in the intelligent domestic equipment investigative range such as intelligent air condition, intelligent refrigerator obtain it is more quick,
Accurately., can be by obvious sunlight, together in regions such as window, door sides but indoor environment is by the illumination effect of complexity
When, the indoor region by light-illuminating is also more wide, and the extraction to characteristics of human body impacts.Indoors, people are mutual
Contact also more frequently, is easily collected within a small range, circumstance of occlusion is even more serious between causing human body, uses RGB etc.
The methods of body templates matching of traditional images and feature learning, can not reach good effect.
Sunlight can be avoided as image acquisition equipment using Kinect and light-illuminating changes and human body blocks mutually
Influence, utilize Kinect obtain comprising depth information image progress human testing.Gray value in depth image only with
Object to equipment camera horizontally or vertically distance dependent, it is unrelated with human body clothing color, the factor such as illumination, have to environment compared with
Good adaptability.Meanwhile the range information that depth image is included causes detection process is similar to carry out target in the 3 d space
Identification, can solve the problems, such as among other color space images that human body blocks well and bring.Depth image possess the above this
A little good performances so that object can be presented than more complete profile in depth information image.
The content of the invention
Part in view of the shortcomings of the prior art, it is an object of the invention to provide one kind to be based on depth image contour
Indoor human body detection method, position of human body that can comprehensively in detection image.
In order to avoid the influence that sunlight and light-illuminating change and human body block mutually, the present invention uses Kinect conducts
Image acquisition equipment, the image comprising depth information obtained using Kinect carry out human testing.Gray scale in depth image
It is worth horizontally or vertically distance dependent only with object to equipment camera, it is unrelated to wear the factors such as color, illumination clothes with human body, to ring
Well adapting to property of border.Meanwhile the range information that depth image is included causes detection process is similar to enter in the 3 d space
Row target identification, can solve the problems, such as among other color space images that human body blocks well and bring.Depth image possesses
The good performance of the above so that object can be presented than more complete profile in depth information image.
Contour refers to the closed curve for connecting into the equal point of height above sea level on topographic map, different each of height above sea level
Individual closed curve vertically scale is projected in the plane of a fixed reference height, and it is obvious contour to be formed level distribution
Line, the digitized representation height above sea level between contour.Contour has the characteristics of following two obvious：
(1) contour of Different Altitude does not intersect.
(2) identical height above sea level mutually non-conterminous a plurality of contour in figure be present.
According to the characteristics of above-mentioned contour, we apply to contour thought among depth information image to represent in scene
The different distance of object and video camera, the height above sea level distributed image that can be formed in scene.It is upright among scene indoors
Human body relative altitude is higher, and closer to pick-up lens, human head and shoulder position is similar to the height on the mountain top and hill-side in topographic map,
The head and shoulder position point of human body in depth map is connected into closed curve and has just obtained human body contour characteristic pattern under current scene.
In summary, contour concept is applied among depth image human testing, can more fast and accurately obtains indoor field
Human body target in scape, accelerate the detection efficiency of smart machine, improve Consumer's Experience.
The purpose of the present invention is achieved through the following technical solutions：
A kind of indoor human body detection method based on depth image contour, its method and step are as follows：
A, depth information image is obtained：Obtain the depth information image of indoor human body；
B, depth image three-dimensional contour is obtained：The depth information of step A images is formed into contour, and profit using software
Contour map is made the contour of different height separate different levels, located as three dimensional stress with the function of Kinect Distance Judgments
It is that the contour curve minimum with imaging device distance is human body head present position in the superiors；
C, top layer three-dimensional contour number of people centralized positioning：Because the head of upright human body is not in contact with normal circumstances
, therefore, the first layer in contour map mutually disjoints；The center for marking each to be in the closing equal pitch contour of first layer
Point, can be to determine a position of human body, and the computational methods of contour central point are as follows：
Wherein, n be the profile coordinate points sum, pxi, pyiFor i-th point of x, y-coordinate, cx, cy are profile central point
X, y-coordinate；The plane coordinates of contour and the coordinate of depth map are one-to-one, and the central point of so contour line profile is sat
Mark can directly corresponds to the number of people center in depth map；
D, depth secondary detection：It is not institute in contour map picture due to the individual difference that the height of people has in itself
Somebody's body head position is in first layer, and this just needs us to extend the level of detection, and then detection is in time high level
Equal pitch contour in number of people position；Due to the relation of contour upright projection, the head contour of same human body is necessarily in
Among the contour that secondary high-rise shoulder is formed, if occurring the human body mark that first layer detects in secondary high-rise contour, say
The bright curve is not shorter individual nose curve；
E, secondary high-rise human body centralized positioning：Use the computational methods in step C, you can it is contour to calculate time closing of high level
The central point of curve, mark time high-rise position of human body；
F, the contour line profile marked and built-in human body head depth image are subjected to silhouette contrast, exclude flase drop；
H, by the calculating of tetra- steps of above step C-F, the human body of different height in depth information image can be marked
Position.
The present invention is to make up human testing false drop rate of the smart machine indoors among environment is high, detection speed is slow etc.
Some are obvious insufficient and propose.Because currently used indoor environment Human Detection passes through image enhaucament, human body mostly
The technologies such as feature templates matching realize that, because indoor environment is complicated, illumination variation is strong, and so as to cause false drop rate high, algorithm takes
It is long, it has not been convenient to which that user uses the smart machine with human testing function.Present invention, avoiding due to being used caused by detection method
Family experience is not good enough, poor for applicability, improves human testing speed, enhancing detection accuracy.
The present invention compared with the prior art, has advantages below and beneficial effect：
Present invention, avoiding due to Consumer's Experience caused by detection method it is not good enough, it is poor for applicability, improve human testing speed,
Strengthen detection accuracy.
Brief description of the drawings
Fig. 1 is the detection method flow chart of the present invention；
Fig. 2 is the depth image three-dimensional contour explanation figure of the present embodiment；
Fig. 3 (A)~Fig. 3 (D) is the Detection results schematic illustration of the present embodiment
Fig. 4 (A)~Fig. 4 (E) is the dense population Detection results schematic illustration of the present embodiment.
Embodiment
The present invention is described in further detail with reference to embodiment：
Embodiment
As shown in Figure 1, Figure 2, it is a kind of based on depth image contour shown in Fig. 3 (A)~Fig. 3 (D), Fig. 4 (A)~Fig. 4 (E)
Indoor human body detection method, in order to avoid sunlight and light-illuminating change and the influence blocked mutually of human body, the present invention uses
For Kinect as image acquisition equipment, the image comprising depth information obtained using Kinect carries out human testing.Depth map
Horizontally or vertically distance dependent of the gray value only with object to equipment camera as in, with human body wear clothes color, illumination etc. because
Element is unrelated, to well adapting to property of environment.Meanwhile the range information that is included of depth image cause detection process similar
Target identification is carried out in 3d space, can solve the problems, such as among other color space images that human body blocks well and bring.It is deep
Degree image possesses the good performance of the above so that object can be presented than more complete profile in depth information image.
Contour refers to the closed curve for connecting into the equal point of height above sea level on topographic map, different each of height above sea level
Individual closed curve vertically scale is projected in the plane of a fixed reference height, and it is obvious contour to be formed level distribution
Line, the digitized representation height above sea level between contour.Contour has the characteristics of following two obvious：
(3) contour of Different Altitude does not intersect.
(4) identical height above sea level mutually non-conterminous a plurality of contour be present.
According to the characteristics of above-mentioned contour, we apply to contour thought among depth information image to represent in scene
The different distance of object and video camera, the height above sea level distributed image that can be formed in scene.It is upright among scene indoors
Human body relative altitude is higher, and closer to pick-up lens, human head and shoulder position is similar to the height on the mountain top and hill-side in topographic map,
The head and shoulder position point of human body in depth map is connected into closed curve and has just obtained human body contour characteristic pattern under current scene.
In summary, contour concept is applied among depth image human testing, can more fast and accurately obtains indoor field
Human body target in scape, accelerate the detection efficiency of smart machine, improve Consumer's Experience.
The purpose of the present invention is achieved through the following technical solutions：
A kind of indoor human body detection method based on depth image contour, its method and step are as follows：
A, depth information image is obtained：Obtain the depth information image of indoor human body；
B, depth image three-dimensional contour is obtained：The depth information of step A images is formed into contour, and profit using software
Contour map is made the contour of different height separate different levels, located as three dimensional stress with the function of Kinect Distance Judgments
It is that the contour curve minimum with imaging device distance is human body head present position in the superiors；
C, top layer three-dimensional contour number of people centralized positioning：Because the head of upright human body is not in contact with normal circumstances
, therefore, the first layer in contour map mutually disjoints；The center for marking each to be in the closing equal pitch contour of first layer
Point, can be to determine a position of human body, and the computational methods of contour central point are as follows：
Wherein, n be the profile coordinate points sum, pxi, pyiFor i-th point of x, y-coordinate, cx, cy are profile central point
X, y-coordinate；The plane coordinates of contour and the coordinate of depth map are one-to-one, and the central point of so contour line profile is sat
Mark can directly corresponds to the number of people center in depth map；
D, depth secondary detection：It is not institute in contour map picture due to the individual difference that the height of people has in itself
Somebody's body head position is in first layer, and this just needs us to extend the level of detection, and then detection is in time high level
Equal pitch contour in number of people position；Due to the relation of contour upright projection, the head contour of same human body is necessarily in
Among the contour that secondary high-rise shoulder is formed, if occurring the human body mark that first layer detects in secondary high-rise contour, say
The bright curve is not shorter individual nose curve；
E, secondary high-rise human body centralized positioning：Use the computational methods in step C, you can it is contour to calculate time closing of high level
The central point of curve, mark time high-rise position of human body；
F, the contour line profile marked and built-in human body head depth image are subjected to silhouette contrast, exclude flase drop；
H, by the calculating of tetra- steps of above step C-F, the human body of different height in depth information image can be marked
Position.The testing result of the present invention is as shown in Fig. 3, accompanying drawing 4.
The present invention is to make up human testing false drop rate of the smart machine indoors among environment is high, detection speed is slow etc.
Some are obvious insufficient and propose.Because currently used indoor environment Human Detection passes through image enhaucament, human body mostly
The technologies such as feature templates matching realize that, because indoor environment is complicated, illumination variation is strong, and so as to cause false drop rate high, algorithm takes
It is long, it has not been convenient to which that user uses the smart machine with human testing function.Present invention, avoiding due to being used caused by detection method
Family experience is not good enough, poor for applicability, improves human testing speed, enhancing detection accuracy.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.
Claims (1)
- A kind of 1. indoor human body detection method based on depth image contour, it is characterised in that：Its method and step is as follows：A, depth information image is obtained：Obtain the depth information image of indoor human body；B, depth image three-dimensional contour is obtained：The depth information of step A images is formed into contour using software, and utilized Contour map as three dimensional stress, is made the contour of different height separate different levels, is in by the function of Kinect Distance Judgments The superiors are that the contour curve minimum with imaging device distance is human body head present position；C, top layer three-dimensional contour number of people centralized positioning：Because the head of upright human body is not in contact with normal circumstances, Therefore, the first layer in contour map mutually disjoints；The central point for marking each to be in the closing equal pitch contour of first layer, just A position of human body can be determined, the computational methods of contour central point are as follows：<mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mi>c</mi> <mi>x</mi> <mo>=</mo> <mfrac> <mn>1</mn> <mi>n</mi> </mfrac> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>px</mi> <mi>i</mi> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mi>c</mi> <mi>y</mi> <mo>=</mo> <mfrac> <mn>1</mn> <mi>n</mi> </mfrac> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <msub> <mi>py</mi> <mi>i</mi> </msub> </mrow> </mtd> </mtr> </mtable> </mfenced>Wherein, n be profile coordinate points sum, pxi, pyiFor i-th point of x, y-coordinate, cx, cy are x, y of profile central point Coordinate；The plane coordinates of contour and the coordinate of depth map are one-to-one, and the center point coordinate of so contour line profile is just The number of people center in depth map can directly be corresponded to；D, depth secondary detection：Due to the individual difference that the height of people has in itself, contour map as in it is not all Human head location is in first layer, and this just needs us to extend the level of detection, and then detection is in time high-rise etc. Number of people position in high curve；Due to the relation of contour upright projection, the head contour of same human body is necessarily high in time Among the contour that layer shoulder is formed, if occurring the human body mark that first layer detects in secondary high-rise contour, illustrate this Curve is not shorter individual nose curve；E, secondary high-rise human body centralized positioning：Use the computational methods in step C, you can calculate time closing equal pitch contour of high level Central point, mark time high-rise position of human body；F, the contour line profile marked and built-in human body head depth image are subjected to silhouette contrast, exclude flase drop；H, by the calculating of tetra- steps of above step C-F, the position of human body of different height in depth information image can be marked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510740255.6A CN105631852B (en) | 2015-11-03 | 2015-11-03 | Indoor human body detection method based on depth image contour |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510740255.6A CN105631852B (en) | 2015-11-03 | 2015-11-03 | Indoor human body detection method based on depth image contour |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105631852A CN105631852A (en) | 2016-06-01 |
| CN105631852B true CN105631852B (en) | 2018-02-06 |
Family
ID=56046738
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510740255.6A Active CN105631852B (en) | 2015-11-03 | 2015-11-03 | Indoor human body detection method based on depth image contour |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105631852B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709444A (en) * | 2016-12-19 | 2017-05-24 | 集美大学 | Binocular infrared photography-based bus passenger flow counting device and method |
| CN106846324B (en) * | 2017-01-16 | 2020-05-01 | 河海大学常州校区 | Irregular object height measuring method based on Kinect |
| CN107392953B (en) * | 2017-09-20 | 2020-11-10 | 四川长虹电器股份有限公司 | Depth image identification method based on contour line |
| CN107588511A (en) * | 2017-09-21 | 2018-01-16 | 四川长虹电器股份有限公司 | Air conditioner energy source management system and method based on contour and HOG human testings |
| CN110501700A (en) * | 2019-08-27 | 2019-11-26 | 四川长虹电器股份有限公司 | A kind of personnel amount method of counting based on millimetre-wave radar |
| CN111079589B (en) * | 2019-12-04 | 2022-09-20 | 常州工业职业技术学院 | Automatic height detection method based on depth camera shooting and height threshold value pixel calibration |
| CN115236661B (en) * | 2022-08-22 | 2023-04-07 | 深圳市海兴科技有限公司 | Human body existence detection method based on microwave induction and microwave inductor |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102122390A (en) * | 2011-01-25 | 2011-07-13 | 于仕琪 | Method for detecting human body based on range image |
| CN103186776A (en) * | 2013-04-03 | 2013-07-03 | 西安电子科技大学 | Human detection method based on multiple features and depth information |
| CN103473571A (en) * | 2013-09-12 | 2013-12-25 | 天津大学 | Human detection method |
| CN103971135A (en) * | 2014-05-05 | 2014-08-06 | 中国民航大学 | Human body target detection method based on head and shoulder depth information features |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040037450A1 (en) * | 2002-08-22 | 2004-02-26 | Bradski Gary R. | Method, apparatus and system for using computer vision to identify facial characteristics |
-
2015
- 2015-11-03 CN CN201510740255.6A patent/CN105631852B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102122390A (en) * | 2011-01-25 | 2011-07-13 | 于仕琪 | Method for detecting human body based on range image |
| CN103186776A (en) * | 2013-04-03 | 2013-07-03 | 西安电子科技大学 | Human detection method based on multiple features and depth information |
| CN103473571A (en) * | 2013-09-12 | 2013-12-25 | 天津大学 | Human detection method |
| CN103971135A (en) * | 2014-05-05 | 2014-08-06 | 中国民航大学 | Human body target detection method based on head and shoulder depth information features |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105631852A (en) | 2016-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105631852B (en) | Indoor human body detection method based on depth image contour | |
| CN108305264B (en) | A kind of unmanned plane precision landing method based on image procossing | |
| CN106647742B (en) | Movement routine method and device for planning | |
| CN106340044B (en) | Join automatic calibration method and caliberating device outside video camera | |
| CN110175576A (en) | A kind of driving vehicle visible detection method of combination laser point cloud data | |
| CN103714541B (en) | Method for identifying and positioning building through mountain body contour area constraint | |
| CN105243664B (en) | A kind of wheeled mobile robot fast-moving target tracking method of view-based access control model | |
| CN110378931A (en) | A kind of pedestrian target motion track acquisition methods and system based on multi-cam | |
| CN103473571B (en) | Human detection method | |
| CN103925927B (en) | A kind of traffic mark localization method based on Vehicular video | |
| CN105608417B (en) | Traffic lights detection method and device | |
| CN107392929B (en) | Intelligent target detection and size measurement method based on human eye vision model | |
| CN110232379A (en) | A kind of vehicle attitude detection method and system | |
| CN104536009A (en) | Laser infrared composite ground building recognition and navigation method | |
| CN103150559A (en) | Kinect three-dimensional depth image-based head identification and tracking method | |
| CN105938619A (en) | Visual odometer realization method based on fusion of RGB and depth information | |
| CN106503170B (en) | It is a kind of based on the image base construction method for blocking dimension | |
| CN108209926A (en) | Human Height measuring system based on depth image | |
| CN106127205A (en) | A kind of recognition methods of the digital instrument image being applicable to indoor track machine people | |
| CN103177247A (en) | Target detection method fused with multi-angle information | |
| CN102915433A (en) | Character combination-based license plate positioning and identifying method | |
| CN106295657A (en) | A kind of method extracting human height's feature during video data structure | |
| CN104143077B (en) | Pedestrian target search method and system based on image | |
| CN106327536A (en) | Neck line measuring method based on sectional point clouds | |
| CN102930524A (en) | Method for detecting heads based on vertically-placed depth cameras |
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 |