CN101996317A - Method and device for identifying markers in human body - Google Patents
Method and device for identifying markers in human body Download PDFInfo
- Publication number
- CN101996317A CN101996317A CN 201010527394 CN201010527394A CN101996317A CN 101996317 A CN101996317 A CN 101996317A CN 201010527394 CN201010527394 CN 201010527394 CN 201010527394 A CN201010527394 A CN 201010527394A CN 101996317 A CN101996317 A CN 101996317A
- Authority
- CN
- China
- Prior art keywords
- label
- human body
- face
- visible images
- infrared light
- 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.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a method for identifying markers in a human body, which comprises the following steps: acquiring a visible light image and an infrared image; extracting imaging regions from the visible light image and the infrared image respectively and identifying a visible light marker and an infrared marker in the imaging regions; matching the visible light marker and the infrared marker, and obtaining the three-dimensional coordinates of the markers; and detecting human body information and eliminating the interferents in the markers. The method and the device for identifying the markers in the human body can eliminate the interferents in the markers of the visible light image and the infrared image according to human body information, thereby accurately identifying the markers in the visible light image and the infrared image and obtaining the dynamic motion track of the human body.
Description
[technical field]
The present invention relates to the visual processes technology, particularly relate to the recognition methods and the device of label in a kind of human body.
[background technology]
In traditional human body in the recognition methods of label, consideration based on real-time, usually adopt two infrared light cameras to obtain the digital video image data of specifying guarded region, the infrared light image that is obtained is handled, and the label on the human body is discerned by the shape facility that extracts in the infrared light image.
Yet, in the operational process of reality, because infrared light image has the advantages that the image-forming information amount is little, interference is many and target is extracted and identification is difficult.For example, in the cap identifying of strip, and indoor fluorescent tube is also elongated, so the imaging of fluorescent tube is very similar to the imaging of cap, is difficult to distinguish cap and fluorescent tube in identifying.
Therefore in traditional people's body tag identifying, if when existing the similar target of above shape, which how to be distinguished is to disturb, and which is that mark on the human body is one and is badly in need of the problem that further solves.
[summary of the invention]
Based on this, be necessary to provide the recognition methods of label in a kind of human body of more accurate identification.
In addition, also be necessary to provide the recognition device of label in a kind of human body of more accurate identification.
The recognition methods of label in a kind of human body comprises the steps: to gather visible images and infrared light image; From described visible images and infrared light image, extract imaging region respectively, and discern visible light label and infrared light label in the described imaging region; Described visible light label and infrared light label are mated, obtain the three-dimensional coordinate of label; Human body information is rejected the chaff interference in the label.
Preferably, described human body information, the process of rejecting the chaff interference in the label is: human body information, judge whether described label contacts with human body, is, confirm that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information comprises at least a in people's face, skin color and the human body contour outline.
Preferably, described human body information behaviour face, described human body information, judge whether described label contacts with human body, be, confirm that then described label is the expection mark, otherwise, the step of then rejecting described label is: carry out people's face and detect in visible images, judge that the imaging of described label and the bee-line between the human face region whether less than default people's face threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information is a skin color, described human body information, judge whether described label contacts with human body, be, confirm that then described label is the expection mark, otherwise, the step of then rejecting described label is: carry out Face Detection in visible images, judge that the imaging of described label and the bee-line between the skin-coloured regions whether less than default colour of skin threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information is a human body contour outline, described human body information, judge whether described label contacts with human body, be, confirm that then described label is the expection mark, otherwise, the step of then rejecting described label is: carry out human body contour outline and detect in visible images, judge that the imaging of described label and the bee-line between the human body contour outline zone whether less than default profile threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
The recognition device of label in a kind of human body comprises at least: acquisition module is used to gather visible images and infrared light image; Picture recognition module is used for extracting imaging region from described visible images and infrared light image respectively, and discerns visible light label and infrared light label in the described imaging region; Matching module is used for described visible light label and infrared light label are mated, and obtains the three-dimensional coordinate of label; Interference cancellation module is used for human body information, rejects the chaff interference in the described label.
Preferably, described interference cancellation module human body information judges whether described label contacts with human body, is, confirms that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information comprises at least a in people's face, skin color and the human body contour outline.
Preferably, described human body information behaviour face, described interference cancellation module comprises people's face processing unit, described people's face processing module is used for carrying out people's face at described visible images and detects, judge that the imaging of described label and the bee-line between the human face region whether less than default people's face threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information is a skin color, described interference cancellation module comprises colour of skin processing unit, described colour of skin processing module is used for carrying out Face Detection at described visible images, judge that the imaging of described label and the bee-line between the skin-coloured regions whether less than default colour of skin threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
Preferably, described human body information comprises human body contour outline, described interference cancellation module comprises the profile processing unit, described profile processing unit is used for carrying out human body contour outline at visible images and detects, judge that the imaging of described label and the bee-line between the human body contour outline zone whether less than default profile threshold value, are, confirm that then described label is the expection mark, otherwise, then reject described label.
Visible images that the recognition methods of label and device will collect in the above-mentioned human body and infrared light image are rejected chaff interference in the label according to human body information, thereby can identify the label in visible images and the infrared light image exactly, and obtain the dynamic moving track of human body by this label.
In the recognition methods of label and the device infrared light image and visible images are carried out the complementation of relative merits in the above-mentioned human body, improved stability and degree of accuracy in people's body tag identifying effectively.
[description of drawings]
Fig. 1 is the process flow diagram of the recognition methods of label in the human body among the embodiment;
Fig. 2 is the process flow diagram of the recognition methods of label in the human body among the embodiment;
Fig. 3 is the synoptic diagram of the recognition device of label in the human body among the embodiment;
Fig. 4 is the detailed block diagram of picture recognition module among the embodiment;
Fig. 5 is the detailed block diagram of matching module among the embodiment;
Fig. 6 is the detailed block diagram of interference cancellation module among the embodiment.
[embodiment]
Fig. 1 shows the method flow of people's body tag identification among the embodiment, comprises the steps:
In step S10, gather visible images and infrared light image.In the present embodiment, obtain visible images and infrared light image in real time, to obtain the multidate information of people's body tag in this visible images and the infrared light image, wherein, the image-forming information amount that infrared light image comprised is less, but image processing speed is very fast, and the image-forming information amount that visible images comprised is big, and image processing speed is slow.
In step S20, from visible images and infrared light image, extract imaging region respectively, and visible light label in the recognition imaging zone and infrared light label.In the present embodiment,, from visible images and infrared light image, extract effective imaging region, so that the label on the human body is discerned by various recognizers according to the type of video image.For example, the label on this human body can be to wear overhead cap, also can be the handle that is held in the hand.
In step S30, infrared light label and visible light label are mated, obtain the three-dimensional coordinate of label.In the present embodiment, respectively infrared light label in the infrared light image and the visible light label in the visible images are mated one by one, thereby under the effect of infrared light image, obtain the three-dimensional coordinate of label.
In step S40, human body information is rejected the chaff interference in the label.In the present embodiment,, be difficult to reject chaff interference in the label,, can reject according to human body information because label contacts with human body based on the recognition methods of shape because existing interference is very many in visible images and the infrared light image." contact " specifically refers to distance between the related zone of the zone at label place and human body information in the preset range content.For example, be positioned over the cap on people's the head, in the process of identification, if this is labeled as cap, then cap and people's appearance connect, and do not link to each other with human body as if mark, and then this is labeled as chaff interference, it need be rejected.Particularly, the process of step S40 is: human body information, and whether the judge mark thing contacts with human body, is, confirms that then this label is the expection mark, otherwise, then reject label.Human body information record the information such as profile profile of human body, comprise at least a in people's face, skin color and the human body contour outline, the expection mark then is the label that needs identification, promptly according to the arbitrary human body information in people's face, skin color and the human body contour outline, whether the imaging of judge mark thing contacts with arbitrary zone in people's face, skin color and the human body contour outline, is to confirm that then label is the expection mark, otherwise, then reject label.
In a preferred embodiment, at least a in human body information behaviour face, skin color and the human body contour outline elaborates the process of rejecting chaff interference by the human body information of people's face, skin color and human body contour outline below.Carry out people's face and detect in visible images, whether the imaging of judge mark thing and the bee-line between the human face region less than default people's face threshold value are, confirm that then this label is the expection mark, otherwise, then reject this label.In the process that people's face in visible images detects, may detect a plurality of human face regions, calculate the imaging of label and the bee-line between all human face regions, and judge that whether this bee-line is less than default people's face threshold value, be, think that then the imaging of label contacts with human face region, otherwise, then think not contact.For example, this people's face threshold value can be 5 pixels.
Carry out Face Detection in visible images, whether the imaging of judge mark thing and the bee-line between the skin-coloured regions less than default colour of skin threshold value, are, confirm that then label is the expection mark, otherwise, then reject this label.In the Face Detection process in visible images, may detect a plurality of area of skin color, calculate the imaging of label and the bee-line between all area of skin color, and judge that whether this bee-line is less than preset threshold value, be, think that then the imaging of label contacts with area of skin color, otherwise, then think not contact.For example, this colour of skin threshold value can be 10 pixels.
Carry out human body contour outline and detect in visible images, whether the bee-line between the imaging of judge mark thing and the human body contour outline zone less than default profile threshold value be, then confirms this label for expecting mark, otherwise, then reject this label.In the human body contour outline testing process of visible images, may detect a plurality of human body contour outlines zone, calculate the imaging of label and the bee-line between all human body contour areas, and judge that whether this bee-line is less than pre-threshold value, be, think that then the imaging of label contacts with the human body contour outline zone, otherwise, then think not contact.For example, this profile threshold value can be 10 pixels.
Fig. 2 shows the recognition methods of label in the human body of an embodiment, comprises the steps:
In step S101, gather infrared light image and visible images respectively.In the present embodiment, gather infrared light image and visible images respectively in real time, so that obtain infrared light image and the event trace that can see people's body tag in the light image simultaneously.The image-forming information amount is little in the infrared light image, image processing speed is fast, and the visible light image information amount is big, and image processing speed is slow, therefore, the complementarity with infrared light image and visible images can effectively improve stability and the identification accuracy of human body label recognition methods under disturbed condition.
In step S102, visible images is retrained, obtain the imaging region of label in visible images, and identification visible light label.In the present embodiment,
Visible images is retrained, the process of obtaining the imaging region of label in visible images specifically: according to outer polar curve constraint, calculate the outer polar curve constraint of label in visible images; According to the time series constraint, label is limited in visible images, obtain the time series constraint; Common factor is got in outer polar curve constraint and time series constraint, obtain the imaging region of label in visible images.By the outer polar curve constraint in the stereoscopic vision, the imaging region of visible images should be that polar curve is a zone at center in addition.
The process of identification visible light label is: discern by recognizer extract proper vector in the imaging region of visible images after, this recognizer can adopt such as support vector machine etc.
In step S103, infrared light image is carried out pre-service, and identification infrared light label.In the present embodiment, infrared light image is carried out carrying out the connected domain detection after the Threshold Segmentation, and, form proper vector each connected domain calculated characteristics.This feature can comprise major axis, minor axis, pixel quantity etc.For example, can utilize the long and short axial length of major axis, area dutycycle, girth of external fitted ellipse square recently to describe shape facility with area, the characteristic of correspondence vector is defined as follows:
Wherein, it is long and minor axis is long that l and w are respectively the major axis of external fitted ellipse, and a and b are respectively the length of imaging region and wide, and S is the area of imaging region, and C is the girth of imaging region.
The identification of infrared light label can be carried out the identification of proper vector by recognizer, and this recognizer can adopt support vector machine etc.
In step S104, obtain the image space sequence matrix according to the infrared light label.In the present embodiment, image space sequence matrix A can obtain by infrared light image being carried out the connected domain detection, is a 2D flag sequence, is designated as { p
1..., p
i, p
M.
In step S105, obtain and the corresponding candidate matches dot matrix of image space sequence matrix according to the visible light label.In the present embodiment, candidate matches dot matrix B can obtain by visible images being carried out the connected domain detection, is a 2D flag sequence, is designated as { q
1..., q
j, q
N.
In step S106,, obtain similarity matrix by image space sequence matrix and candidate matches dot matrix.In the present embodiment, similarity matrix C={c
I, j, c wherein
I, jBe p
iAnd q
jSimilarity.
In step S107, according to default minimum similarity, the value less than minimum similarity in the similarity matrix is set to zero.In the present embodiment, minimum similarity is an empirical value, can set flexibly according to actual conditions, and for example, minimum similarity can be 0.2, with similarity matrix C={c
I, jIn be set to zero less than the value of minimum similarity, go exchanges, row exchange, obtain
Form.
In step S108, by similarity matrix, calculate intermediate parameters and total similarity, if intermediate parameters is 1, then the pairing visible light label of pairing infrared light label of image space sequence and candidate matches dot matrix mates.In the present embodiment, pass through formula
Calculate intermediate parameters x
IjAnd total similarity F, wherein,
I=1,2 ... M,
J=1,2 ... N, x
IjBeing 0 or 1, is the matrix of M*N.
In step S109, human body information, whether the judge mark thing contacts with human body, be then to enter step S110, otherwise, then enter step S111.In the present embodiment, for the label in the human body, for example, if label is a cap, then people's face should be below cap, if label is a handle, then handle should be near hand, and therefore label on one's body should, can utilize human body information to come whether the judge mark thing is chaff interference in people's profile.According to human body information, the judge mark thing whether with process that the people contacts in, if label contacts with human body, then this label is the expection mark, if label does not contact with human body, then this label is a chaff interference, should be rejected.Particularly, for elaborating, be that example proves absolutely this judgement content below with the cap according to the human body information judge mark thing process of chaff interference whether.
For cap, because cap is to be worn on people's the head, its imaging should be on the border of human body contour outline, should there be a slice area of skin color its imaging below, being people's face, therefore, is that image object comes to judge by following three kinds of modes whether this image object is cap with the imaging region of cap.
Among one embodiment, end user's face judges that image object is a chaff interference, or cap, promptly use Adaboost algorithm (Adaptive Boosting, machine learning method) detects people's face, if under the imaging region of cap, detected people's face, confirm on this basis that then this image object is a cap.
Among another embodiment, end user's profile judges that image object is chaff interference or cap.The realization approach that people's profile detects can be divided into two classes, and the first kind is based on the method for partial model, takes the approach from parts to integral body.For example, human body is divided into face, left arm, right arm and four parts of leg, train the detecting device of these four parts then respectively, at last detect whole human body according to the geometrical constraint between the parts, in addition, also can utilize the position histogram feature to set up a upper limbs detecting device based on the parts of people's face, head and shoulder.Second class is based on the method for single detection window, judges at yardstick and locational space application class device whether all images are cap.For example, utilize Haar (Ha Er) feature and polynomial expression support vector machine (the support vector machine that revised, be called for short SVM) as the human body detector of sorter, in addition, also can utilize Haar feature and motion feature Adaboosting sorter, construct a video human face detecting device fast in conjunction with cascade.The profile of human body detects can also use Harr human body contour outline sorter.
Among other embodiment, use the colour of skin to judge that image object is chaff interference or cap.Because biological and colour of skin Uniformity of Distribution physically, although people's the colour of skin is different because of the difference of ethnic group, present different colors, but after having got rid of the influence factors to the colour of skin such as brightness, environment, the tone of skin is a basically identical, therefore can utilize colouring information to provide favourable evidence for skin detection.With the YIQ space is example, and Y is the gray-scale value (Gray value) of image, and I and Q then are meant tone (Chrominance), promptly describe the attribute of image color and saturation degree.In the YIQ space, the monochrome information of Y component representative image, two components of I, Q then carry colouring information, the change color of I component representative from orange to cyan, Q component is then represented from purple to yellowish green change color.Use complexion model to carry out binaryzation in the YIQ space, to obtain the binary map of the colour of skin.Complexion model is
Wherein r and q represent the scope of I component and Q component respectively, and m represents the average of I component, and I component and Q component then carry colouring information respectively, the change color of I component representative from orange to cyan, and Q component is then represented from purple to yellowish green change color.Usually visible images that collects and infrared light image are the image of rgb format, and each pixel is all comprising R, G and three components of B, and the conversion from rgb space to the YIQ space can be to be expressed in matrix as:
Pixel p in the image transforms to the YIQ space with it from rgb space, if satisfy inequality
Then pixel p is a skin pixel, otherwise is non-skin pixel.Get m=55 in a preferred embodiment, r=40, best results during q=20.
In step S110, confirm that label is the expection mark.
In step S111, reject label.
Fig. 3 shows among the embodiment recognition device of label in the human body, and this device comprises at least:
Fig. 4 shows the picture recognition module of an embodiment, and picture recognition module 20 comprises at least:
Acquiring unit 202 is used for obtaining imaging region from the constraint, and the visible light label in the recognition imaging zone.In the present embodiment, acquiring unit 202 is got common factor with outer polar curve constraint and time series constraint, obtain the imaging region of label in visible images, and discern by recognizer after in the imaging region of visible images, extracting proper vector, this recognizer can adopt the support vector machine scheduling algorithm.
Wherein, it is long and minor axis is long that l and w are respectively the major axis of external fitted ellipse, and a and b are respectively the length of imaging region and wide, and S is the area of imaging region, and C is the girth of imaging region.
Infrared processing unit can carry out the identification of proper vector by recognizer to the identification of infrared light label, and this recognizer can adopt support vector machine etc.
Fig. 5 shows the matching module 30 of an embodiment, and this matching module 30 comprises extraction unit 301 and similarity computing unit 302, wherein:
Extraction unit 301 is used for obtaining the image space sequence matrix according to the infrared light mark, and obtains and the corresponding candidate matches dot matrix of image space sequence matrix according to the visible light mark.In the present embodiment, extraction unit 301 obtains image space sequence matrix A by infrared light image being carried out the connected domain detection, visible images is carried out connected domain detect and obtain candidate matches dot matrix B.This image space sequence matrix A is the 2D flag sequence, is designated as { q
1..., q
j, q
N, candidate matches dot matrix B is the 2D sequence, is designated as { q
1..., q
j, q
N.
Similarity computing unit 302, be used for by image space sequence matrix and candidate matches dot matrix, obtain similarity matrix, according to default minimum similarity, value less than minimum similarity in the similarity matrix is set to zero, and calculate intermediate parameters and total similarity, and if intermediate parameters is 1, then pairing infrared light mark of this image space sequence and the pairing visible light indicia matched of candidate matches dot matrix.In the present embodiment, similarity matrix C={c
I, j, c wherein
I, jBe p
iAnd q
jSimilarity, similarity computing unit 302 is according to default minimum similarity, with similarity matrix C={c
I, jIn be set to zero less than the value of minimum similarity, go exchanges, row exchange, obtain
Form, and pass through formula
Calculate intermediate parameters xij and total similarity F, wherein,
I=1,2 ... M,
J=1,2 ... N, x
IjBeing 0 or 1, is the matrix of M*N.
Fig. 6 shows the interference cancellation module of an embodiment, because human body information comprises at least a in people's face, skin color and the human body contour outline, therefore according to actual needs, this interference cancellation module 40 has comprised at least a in people's face processing unit 410, colour of skin processing unit 430 and the profile processing unit 450, wherein:
People's face processing unit 410 is used for carrying out people's face at visible images and detects, and whether the imaging of judge mark thing and the bee-line between the human face region less than default people's face threshold value are, confirms that then this label is the expection mark, otherwise, then reject this label.In people's face processing unit 410 carries out process that people's face of visible images detects, may detect a plurality of human face regions, calculate the imaging and the bee-line between all human face regions of label and judge that whether this bee-line is less than default people's face threshold value, be, think that then the imaging of label contacts with human face region, otherwise, then think not contact.For example, this people's face threshold value can be 5 pixels.
Colour of skin processing unit 430 is used for carrying out Face Detection at visible images, and whether the imaging of judge mark thing and the bee-line between the skin-coloured regions be less than default colour of skin threshold value, be, confirm that then label is the expection mark, otherwise, this label then rejected.Colour of skin processing unit 430 is in the Face Detection process of visible images, may detect a plurality of area of skin color, calculate the imaging of label and the bee-line between all area of skin color, and judge that whether this bee-line is less than preset threshold value, be, think that then the imaging of label contacts with area of skin color, otherwise, then think not contact.For example, this colour of skin threshold value can be 10 pixels.
Visible images that the recognition methods of label and device will collect in the above-mentioned human body and infrared light image are rejected chaff interference in the label according to human body information, thereby identify the label in visible images and the infrared light image exactly, and obtain the dynamic moving track of human body by this label.
In the recognition methods of label and the device infrared light image and visible images are carried out the complementation of relative merits in the above-mentioned human body, improved stability and degree of accuracy in people's body tag identifying effectively.
The above embodiment has only expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.
Claims (12)
1. the recognition methods of label in the human body comprises the steps:
Gather visible images and infrared light image;
From described visible images and infrared light image, extract imaging region respectively, and discern visible light label and infrared light label in the described imaging region;
Described visible light label and infrared light label are mated, obtain the three-dimensional coordinate of label;
Human body information is rejected the chaff interference in the label.
2. the recognition methods of label is characterized in that in the human body according to claim 1, described human body information, and the process of rejecting the chaff interference in the label is:
Human body information judges whether described label contacts with human body, is, confirms that then described label is the expection mark, otherwise, then reject described label.
3. the recognition methods of label is characterized in that in the human body according to claim 2, and described human body information comprises at least a in people's face, skin color and the human body contour outline.
4. the recognition methods of label in the human body according to claim 3, it is characterized in that, described human body information behaviour face, described human body information, judge whether described label contacts with human body, is, confirm that then described label is the expection mark, otherwise the step of then rejecting described label is:
In visible images, carry out people's face and detect, judge that the imaging of described label and the bee-line between the human face region whether less than default people's face threshold value, are, confirm that then described label is the expection mark, otherwise, described label then rejected.
5. the recognition methods of label in the human body according to claim 3, it is characterized in that, described human body information is a skin color, described human body information, judge whether described label contacts with human body, is, confirm that then described label is the expection mark, otherwise the step of then rejecting described label is:
In visible images, carry out Face Detection, judge that the imaging of described label and the bee-line between the skin-coloured regions whether less than default colour of skin threshold value, are, confirm that then described label is the expection mark, otherwise, described label then rejected.
6. the recognition methods of label in the human body according to claim 3, it is characterized in that, described human body information is a human body contour outline, described human body information, judge whether described label contacts with human body, is, confirm that then described label is the expection mark, otherwise the step of then rejecting described label is:
Carry out human body contour outline and detect in visible images, whether the imaging of judging described label and the bee-line between the human body contour outline zone less than default profile threshold value be, then confirms described label for expecting mark, otherwise, then reject described label.
7. the recognition device of label in the human body is characterized in that, comprises at least:
Acquisition module is used to gather visible images and infrared light image;
Picture recognition module is used for extracting imaging region from described visible images and infrared light image respectively, and discerns visible light label and infrared light label in the described imaging region;
Matching module is used for described visible light label and infrared light label are mated, and obtains the three-dimensional coordinate of label;
Interference cancellation module is used for human body information, rejects the chaff interference in the described label.
8. the recognition device of label is characterized in that in the human body according to claim 7, described interference cancellation module human body information, judge whether described label contacts with human body, is, confirm that then described label is the expection mark, otherwise, then reject described label.
9. the recognition device of label is characterized in that in the human body according to claim 8, and described human body information comprises at least a in people's face, skin color and the human body contour outline.
10. the recognition device of label in the human body according to claim 9, it is characterized in that, described human body information behaviour face, described interference cancellation module comprises people's face processing unit, and described people's face processing module is used for carrying out people's face at described visible images and detects, and judges that whether the imaging of described label and the bee-line between the human face region are less than default people's face threshold value, be, confirm that then described label is the expection mark, otherwise, described label then rejected.
11. the recognition device of label in the human body according to claim 9, it is characterized in that, described human body information is a skin color, described interference cancellation module comprises colour of skin processing unit, and described colour of skin processing module is used for carrying out Face Detection at described visible images, judges that whether the imaging of described label and the bee-line between the skin-coloured regions are less than default colour of skin threshold value, be, confirm that then described label is the expection mark, otherwise, described label then rejected.
12. the recognition device of label in the human body according to claim 9, it is characterized in that, described human body information comprises human body contour outline, described interference cancellation module comprises the profile processing unit, and described profile processing unit is used for carrying out human body contour outline at visible images and detects, and whether the imaging of judging described label and the bee-line between the human body contour outline zone be less than the profile threshold value of presetting, be, confirm that then described label is the expection mark, otherwise, described label then rejected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105273948A CN101996317B (en) | 2010-11-01 | 2010-11-01 | Method and device for identifying markers in human body |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010105273948A CN101996317B (en) | 2010-11-01 | 2010-11-01 | Method and device for identifying markers in human body |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101996317A true CN101996317A (en) | 2011-03-30 |
| CN101996317B CN101996317B (en) | 2012-11-21 |
Family
ID=43786455
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010105273948A Active CN101996317B (en) | 2010-11-01 | 2010-11-01 | Method and device for identifying markers in human body |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101996317B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102629315A (en) * | 2012-02-29 | 2012-08-08 | 北京无线电计量测试研究所 | Automatic detection and identification apparatus of concealed object |
| WO2014040559A1 (en) * | 2012-09-14 | 2014-03-20 | 华为技术有限公司 | Scene recognition method and device |
| CN105391917A (en) * | 2015-11-05 | 2016-03-09 | 清华大学 | Interferent-detection-based video optimization method |
| CN106471523A (en) * | 2014-06-30 | 2017-03-01 | 微软技术许可有限责任公司 | Colour code using infrared imaging |
| CN109078270A (en) * | 2018-06-29 | 2018-12-25 | 天津市宝坻区人民医院 | A kind of auxiliary infrared positioning apparatus and method based on radiotherapy |
| CN109685078A (en) * | 2018-12-17 | 2019-04-26 | 浙江大学 | Infrared image recognition based on automatic marking |
| CN110411570A (en) * | 2019-06-28 | 2019-11-05 | 武汉高德智感科技有限公司 | Infrared human body temperature screening method based on human testing and human body tracking technology |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1932882A (en) * | 2006-10-19 | 2007-03-21 | 上海交通大学 | Infared and visible light sequential image feature level fusing method based on target detection |
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN101739550A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Method and system for detecting moving objects |
-
2010
- 2010-11-01 CN CN2010105273948A patent/CN101996317B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1932882A (en) * | 2006-10-19 | 2007-03-21 | 上海交通大学 | Infared and visible light sequential image feature level fusing method based on target detection |
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN101739550A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Method and system for detecting moving objects |
Non-Patent Citations (1)
| Title |
|---|
| 《Proc.SPIE》 19981231 O.Rockinger等 Pix-level image fusion:the case of image sequences , * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102629315A (en) * | 2012-02-29 | 2012-08-08 | 北京无线电计量测试研究所 | Automatic detection and identification apparatus of concealed object |
| US9465992B2 (en) | 2012-09-14 | 2016-10-11 | Huawei Technologies Co., Ltd. | Scene recognition method and apparatus |
| WO2014040559A1 (en) * | 2012-09-14 | 2014-03-20 | 华为技术有限公司 | Scene recognition method and device |
| CN106471523B (en) * | 2014-06-30 | 2020-03-03 | 微软技术许可有限责任公司 | Color identification using infrared imaging |
| CN106471523A (en) * | 2014-06-30 | 2017-03-01 | 微软技术许可有限责任公司 | Colour code using infrared imaging |
| US10936900B2 (en) | 2014-06-30 | 2021-03-02 | Microsoft Technology Licensing, Llc | Color identification using infrared imaging |
| CN105391917B (en) * | 2015-11-05 | 2018-10-16 | 清华大学 | Method for optimizing video based on interference analyte detection |
| CN105391917A (en) * | 2015-11-05 | 2016-03-09 | 清华大学 | Interferent-detection-based video optimization method |
| CN109078270A (en) * | 2018-06-29 | 2018-12-25 | 天津市宝坻区人民医院 | A kind of auxiliary infrared positioning apparatus and method based on radiotherapy |
| CN109685078A (en) * | 2018-12-17 | 2019-04-26 | 浙江大学 | Infrared image recognition based on automatic marking |
| CN109685078B (en) * | 2018-12-17 | 2022-04-05 | 浙江大学 | Infrared image identification method based on automatic annotation |
| CN110411570A (en) * | 2019-06-28 | 2019-11-05 | 武汉高德智感科技有限公司 | Infrared human body temperature screening method based on human testing and human body tracking technology |
| CN110411570B (en) * | 2019-06-28 | 2020-08-28 | 武汉高德智感科技有限公司 | Infrared human body temperature screening method based on human body detection and human body tracking technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101996317B (en) | 2012-11-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101996317B (en) | Method and device for identifying markers in human body | |
| CN107578035B (en) | Human body contour extraction method based on super-pixel-multi-color space | |
| CN106682601B (en) | A kind of driver's violation call detection method based on multidimensional information Fusion Features | |
| CN108319973B (en) | Detection method for citrus fruits on tree | |
| CN108520226B (en) | Pedestrian re-identification method based on body decomposition and significance detection | |
| CN105160317B (en) | One kind being based on area dividing pedestrian gender identification method | |
| US6611613B1 (en) | Apparatus and method for detecting speaking person's eyes and face | |
| CN105844242A (en) | Method for detecting skin color in image | |
| CN108921119B (en) | Real-time obstacle detection and classification method | |
| CN105389581B (en) | A kind of rice germ plumule integrity degree intelligent identifying system and its recognition methods | |
| Tsai et al. | Road sign detection using eigen colour | |
| CN103839042B (en) | Face identification method and face identification system | |
| Skodras et al. | An unconstrained method for lip detection in color images | |
| CN108629319B (en) | Image detection method and system | |
| CN106548139B (en) | A kind of pedestrian's recognition methods again | |
| CN112232332B (en) | Non-contact palm detection method based on video sequence | |
| CN105787481B (en) | A kind of object detection method and its application based on the potential regional analysis of Objective | |
| CN102096823A (en) | Face detection method based on Gaussian model and minimum mean-square deviation | |
| CN106503644B (en) | Glasses attribute detection method based on edge projection and color characteristic | |
| CN111967363B (en) | Emotion prediction method based on micro-expression recognition and eye movement tracking | |
| CN105894503A (en) | Method for restoring Kinect plant color and depth detection images | |
| CN109271932A (en) | Pedestrian based on color-match recognition methods again | |
| CN103034838A (en) | Special vehicle instrument type identification and calibration method based on image characteristics | |
| CN105844245A (en) | Fake face detecting method and system for realizing same | |
| CN107392151A (en) | Face image various dimensions emotion judgement system and method based on neutral net |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |