CN113743276A - Method for judging human body part where target object is located in human body gray level image - Google Patents
Method for judging human body part where target object is located in human body gray level image Download PDFInfo
- Publication number
- CN113743276A CN113743276A CN202111002803.7A CN202111002803A CN113743276A CN 113743276 A CN113743276 A CN 113743276A CN 202111002803 A CN202111002803 A CN 202111002803A CN 113743276 A CN113743276 A CN 113743276A
- Authority
- CN
- China
- Prior art keywords
- human body
- algorithm
- dangerous goods
- body part
- target object
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 15
- 238000001514 detection method Methods 0.000 claims abstract description 17
- 238000007689 inspection Methods 0.000 claims abstract description 10
- 210000000746 body region Anatomy 0.000 claims abstract description 4
- 230000002093 peripheral effect Effects 0.000 claims description 4
- 210000003414 extremity Anatomy 0.000 description 10
- 231100001261 hazardous Toxicity 0.000 description 8
- 238000003384 imaging method Methods 0.000 description 4
- 210000000689 upper leg Anatomy 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 210000001364 upper extremity Anatomy 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a method for judging a human body part where a target object is located in a human body gray image, and when a dangerous article is carried by a security check person, the security check device can quickly inform the security check person of the human body part where the dangerous article is located, so as to help the security check person to check. The method is characterized in that: 1, firstly, capturing a gray image of a human body in a security inspection area; 2, retraining a proper model on the gray-scale image data by utilizing an algorithm network structure according to a human body posture key point algorithm of visible light: a, determining required human body key points according to actual application requirements and human body regions needing to be divided; b, based on the human body key points, marking out approximate human body parts through an algorithm; and 3, sorting the dangerous goods detection target frames based on the coincidence degree of the detected dangerous goods detection target frames and each human body part in the upper graph, wherein the coincidence degree is the highest and exceeds a set threshold value, and the dangerous goods detection target frames are the human body parts where the dangerous goods are located.
Description
Technical Field
The invention relates to the technical field of security check, in particular to a method for judging a human body part where a target object is located in a human body gray level image.
Background
A plurality of novel human body security inspection technologies are widely applied, and the human body is imaged by infrared, terahertz waves, laser radars and the like, wherein the imaging is different from a common visible light camera, and the imaging is gray imaging essentially. In the existing security check application, some target detection algorithms can detect dangerous goods which are carried in a hidden way by a human body, but certain algorithms are needed to judge the position of the dangerous goods in the human body for privacy or quick security check.
At present, the mature algorithm of the gray level image is less applied, and no direct visible light algorithm can be used for reference, so a new scheme needs to be provided to solve the application problem.
Disclosure of Invention
In order to solve the problems, the invention provides a method for judging the human body part where a target object is located in a human body gray scale image.
A method for judging a human body part where a target object is located in a human body gray level image is characterized by comprising the following steps:
1, firstly, capturing a gray image of a human body in a security inspection area;
2, retraining a proper model on the gray-scale image data by utilizing an algorithm network structure according to a human body posture key point algorithm of visible light:
a, determining required human body key points according to actual application requirements and human body regions needing to be divided;
b, based on the human body key points, marking out approximate human body parts through an algorithm;
and 3, sorting the dangerous goods detection target frames based on the coincidence degree of the detected dangerous goods detection target frames and each human body part in the upper graph, wherein the coincidence degree is the highest and exceeds a set threshold value, and the dangerous goods detection target frames are the human body parts where the dangerous goods are located.
It is further characterized in that:
the algorithm of the step a comprises two algorithms, wherein the first algorithm is a rectangular area formed by symmetrically expanding key point connecting lines in a left-right mode, and the second algorithm is a polygonal area formed by enclosing the existing key points and supplementary points calculated based on the existing key points;
for the four limbs and the head, corresponding to the first algorithm, when the corresponding key point connecting line is positioned near the central axis of each of the four limbs or the head, the four limbs or the head are expanded left and right on the basis of the connecting line to form a rectangular area, the expansion sizes of the four limbs or the head are different from each other, and each area has a corresponding experience value;
for the torso part, corresponding to the second algorithm, and corresponding key points are marked near the peripheral contour of the torso, some supplementary points are properly calculated according to corresponding requirements, and then corresponding polygonal areas are enclosed by using the connecting lines of the points;
the algorithm of the step a is not limited to the two algorithms, and proper key points are selected according to different scene requirements, the key points are different, the design ideas of the algorithm are different, and the design of the algorithm and the selection of the key points are dense and inseparable and mutually influence each other;
the coincidence degree in the step 3 is specifically the ratio of the coincidence area to the area of the dangerous goods detection frame.
After adopting above-mentioned technical scheme, carry the target object by the security check personnel and carry out grey level image and catch through the security check equipment, human key point is caught according to the human gesture key point algorithm of the gesture imitation visible light of grey level image, later based on above-mentioned human key point, through the algorithm, mark off approximate human position, then detect the coincidence degree of target frame and everybody position in the last picture based on the hazardous articles that have detected, sort, the coincidence degree is the highest, and exceed and set for the threshold value, for the human position that the hazardous articles locates, when being carried the hazardous articles by the security check personnel, the security check equipment can inform the human position that the hazardous articles locates fast, help the inspection personnel to check.
Drawings
FIG. 1 is a gray scale human imaging in an embodiment;
FIG. 2 illustrates key points of a human body obtained according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a human body part being divided according to an embodiment of the present invention;
FIG. 4 is a final result presentation of the method of the present invention.
Detailed Description
A method for judging a human body part where a target object is located in a human body gray level image comprises the following specific steps:
1, firstly, capturing a gray image of a human body in a security inspection area;
2, retraining a proper model on the gray-scale image data by utilizing an algorithm network structure according to a human body posture key point algorithm of visible light:
a, determining required human body key points according to actual application requirements and human body regions needing to be divided;
b, based on the human body key points, marking out approximate human body parts through an algorithm;
and 3, sorting the dangerous goods detection target frames based on the coincidence degree of the detected dangerous goods detection target frames and each human body part in the upper graph, wherein the coincidence degree is the highest and exceeds a set threshold value, and the dangerous goods detection target frames are the human body parts where the dangerous goods are located.
The algorithm of the step a comprises two algorithms, wherein the first algorithm is a rectangular area formed by symmetrically expanding key point connecting lines in a left-right mode, and the second algorithm is a polygonal area formed by enclosing the existing key points and supplementary points calculated based on the existing key points;
for the four limbs and the head, corresponding to the first algorithm, when the corresponding key point connecting line is positioned near the central axis of each of the four limbs or the head, the four limbs or the head are expanded left and right on the basis of the connecting line to form a rectangular area, the expansion sizes of the four limbs or the head are different from each other, and each area has a corresponding experience value;
for the torso portion, which corresponds to the second algorithm, the corresponding key points are marked near the peripheral contour of the torso, some supplementary points are calculated appropriately according to the corresponding requirements, and then the points are connected to form the corresponding polygonal area.
The algorithm of the step a is not limited to the two algorithms, and proper key points are selected according to different scene requirements, the key points are different, the design ideas of the algorithm are different, and the design of the algorithm and the selection of the key points are dense and inseparable and mutually influence each other;
the coincidence degree in the step 3 is specifically the ratio of the coincidence area to the area of the dangerous goods detection frame.
In specific implementation, 1, firstly, a standing gray-scale image of a human body in a security inspection area is captured (see fig. 1);
2, according to the key point algorithm of the human body vertical standing posture of the visible light, finding out 20 key points of the human body (see fig. 2), namely two heads, three limbs and eight trunk parts, wherein two coincident key points exist on two sides of the upper positions of the trunk part and the upper limbs, the coincident key points are actually one point, and retraining a proper model on gray scale image data by utilizing an algorithm network structure:
when the corresponding key point connecting line is positioned near the central axis of each of the four limbs or the head, the connecting line is expanded left and right to form a rectangular area, the expansion size of each area is different, and each area has a corresponding experience value;
for the trunk part, corresponding key points are marked near the peripheral outline of the trunk, supplementary points are calculated and divided appropriately according to corresponding requirements, and then corresponding polygonal areas are enclosed by connecting lines of the points to divide the trunk into six parts; in specific implementation, the human body in the standing posture is divided into 15 parts in total, namely a head, a left upper arm, a left lower arm, a right upper arm, a right lower arm, a trunk, a left thigh, a left shank, a right thigh and a right shank (see figure 3).
3 based on the contact ratio of the detected dangerous article detection target frame and each human body part in the previous figure, sequencing is carried out, a dangerous article target exists at the right side position of the middle part and the right shank part of the trunk, the algorithm result is visually displayed (see figure 4), during actual application, the number or the name of the human body part where the target object is located is finally output by an algorithm interface corresponding to the whole method, a corresponding display is carried out after the corresponding number or the name is obtained by a security inspection system, and the display is displayed on a display of the system, so that security inspection personnel can conveniently and rapidly and accurately carry out inspection.
The working principle is as follows: carry the target object by the security check personnel and carry out grey level image and catch through security check equipment, human gesture key point algorithm according to grey level image's gesture imitation visible light catches human key point, later based on above-mentioned human key point, through the algorithm, mark off approximate human position, then detect the contact degree of each human body position in target frame and the last picture based on the hazardous articles that have detected, sort, contact degree is the highest, and exceed and set for the threshold value, for the human position that the hazardous articles locates, when being carried the hazardous articles by the security check personnel, security check equipment can inform security check personnel hazardous articles place human position fast, help security check personnel to check.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (5)
1. A method for judging a human body part where a target object is located in a human body gray level image is characterized by comprising the following steps:
1, firstly, capturing a gray image of a human body in a security inspection area;
2, retraining a proper model on the gray-scale image data by utilizing an algorithm network structure according to a human body posture key point algorithm of visible light:
a, determining required human body key points according to actual application requirements and human body regions needing to be divided;
b, based on the human body key points, marking out approximate human body parts through an algorithm;
and 3, sorting the dangerous goods detection target frames based on the coincidence degree of the detected dangerous goods detection target frames and each human body part in the upper graph, wherein the coincidence degree is the highest and exceeds a set threshold value, and the dangerous goods detection target frames are the human body parts where the dangerous goods are located.
2. The method for determining the human body part where the target object is located in the human body gray scale image as claimed in claim 1, wherein: the algorithm of the step a comprises two algorithms, wherein the first algorithm is a rectangular area formed by symmetrically expanding key point connecting lines in a left-right mode, and the second algorithm is a polygonal area formed by enclosing the existing key points and supplementary points calculated based on the existing key points.
3. The method for determining the human body part where the target object is located in the human body gray scale image as claimed in claim 1, wherein: for the four limbs and the head, which correspond to the first algorithm, when the corresponding key point connecting line is located near the central axis of each of the four limbs or the head, the four limbs or the head are expanded left and right on the basis of the connecting line to form a rectangular area, the expansion sizes of the four limbs or the head are different from each other, and each area has a corresponding empirical value.
4. The method for determining the human body part where the target object is located in the human body gray scale image as claimed in claim 1, wherein: for the torso portion, which corresponds to the second algorithm, the corresponding key points are marked near the peripheral contour of the torso, some supplementary points are calculated appropriately according to the corresponding requirements, and then the points are connected to form the corresponding polygonal area.
5. The method for determining the human body part where the target object is located in the human body gray scale image as claimed in claim 1, wherein: the coincidence degree in the step 3 is specifically the ratio of the coincidence area to the area of the dangerous goods detection frame.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111002803.7A CN113743276A (en) | 2021-08-30 | 2021-08-30 | Method for judging human body part where target object is located in human body gray level image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111002803.7A CN113743276A (en) | 2021-08-30 | 2021-08-30 | Method for judging human body part where target object is located in human body gray level image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113743276A true CN113743276A (en) | 2021-12-03 |
Family
ID=78733877
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111002803.7A Pending CN113743276A (en) | 2021-08-30 | 2021-08-30 | Method for judging human body part where target object is located in human body gray level image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113743276A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108956526A (en) * | 2018-06-22 | 2018-12-07 | 西安天和防务技术股份有限公司 | A kind of passive type Terahertz hazardous material detection device, detection method and its application |
| CN109799544A (en) * | 2018-12-28 | 2019-05-24 | 深圳市华讯方舟太赫兹科技有限公司 | Intelligent detecting method, device and storage device applied to millimeter wave safety check instrument |
| US20190370537A1 (en) * | 2018-05-29 | 2019-12-05 | Umbo Cv Inc. | Keypoint detection to highlight subjects of interest |
| CN112560741A (en) * | 2020-12-23 | 2021-03-26 | 中国石油大学(华东) | Safety wearing detection method based on human body key points |
-
2021
- 2021-08-30 CN CN202111002803.7A patent/CN113743276A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190370537A1 (en) * | 2018-05-29 | 2019-12-05 | Umbo Cv Inc. | Keypoint detection to highlight subjects of interest |
| CN108956526A (en) * | 2018-06-22 | 2018-12-07 | 西安天和防务技术股份有限公司 | A kind of passive type Terahertz hazardous material detection device, detection method and its application |
| CN109799544A (en) * | 2018-12-28 | 2019-05-24 | 深圳市华讯方舟太赫兹科技有限公司 | Intelligent detecting method, device and storage device applied to millimeter wave safety check instrument |
| CN112560741A (en) * | 2020-12-23 | 2021-03-26 | 中国石油大学(华东) | Safety wearing detection method based on human body key points |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10417775B2 (en) | Method for implementing human skeleton tracking system based on depth data | |
| CN102657532B (en) | Height measuring method and device based on body posture identification | |
| US9898651B2 (en) | Upper-body skeleton extraction from depth maps | |
| JP5837508B2 (en) | Posture state estimation apparatus and posture state estimation method | |
| CN104751146B (en) | A kind of indoor human body detection method based on 3D point cloud image | |
| CN110147738B (en) | Driver fatigue monitoring and early warning method and system | |
| US20160252346A1 (en) | Determining three-dimensional information from projections or placement of two-dimensional patterns | |
| CN106991654A (en) | Human body beautification method and apparatus and electronic installation based on depth | |
| CN109506628A (en) | Object distance measuring method under a kind of truck environment based on deep learning | |
| JP2004050390A (en) | Work taking out device | |
| GB2584400A (en) | Processing captured images | |
| CN107766819A (en) | A kind of video monitoring system and its real-time gait recognition methods | |
| CN106960473B (en) | behavior perception system and method | |
| JP2019096113A (en) | Processing device, method and program relating to keypoint data | |
| CN106373336A (en) | Fall detection method and device | |
| CN112185514A (en) | Rehabilitation training effect evaluation system based on action recognition | |
| KR20220024494A (en) | Method and system for human monocular depth estimation | |
| CN111222379A (en) | Hand detection method and device | |
| EP4075388A1 (en) | Computing device and method for generating realistic synthetic image data | |
| AU2018430831B2 (en) | Determining method, determining program, and information processing apparatus | |
| CN110222614A (en) | A kind of method and apparatus that identification pedestrian falls | |
| CN113743276A (en) | Method for judging human body part where target object is located in human body gray level image | |
| CN109740458B (en) | Method and system for measuring physical characteristics based on video processing | |
| CN116229235A (en) | Human body posture estimation network model and estimation method based on thermal imaging | |
| JPH0273471A (en) | Estimating method for three-dimensional form |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |