CN107103617A - The recognition methods of safety cap wearing state and system based on optical flow method - Google Patents
The recognition methods of safety cap wearing state and system based on optical flow method Download PDFInfo
- Publication number
- CN107103617A CN107103617A CN201710188270.3A CN201710188270A CN107103617A CN 107103617 A CN107103617 A CN 107103617A CN 201710188270 A CN201710188270 A CN 201710188270A CN 107103617 A CN107103617 A CN 107103617A
- Authority
- CN
- China
- Prior art keywords
- head
- safety cap
- shoulder
- extraction
- wearing state
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/269—Analysis of motion using gradient-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- 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
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- 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)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses the safety cap wearing state recognition methods based on optical flow method and system, the method comprising the steps of:S1, the monitor video for obtaining job site in real time, and obtain the extraction frame of user preset;S2, according to default extraction frame, after the picture frame of extract real-time monitor video, Difference Calculation is based on using optical flow method and carries out movement human Objective extraction;S3, the movement human target progress head and shoulder detection to extracting acquisition, extract the head shoulder images in obtaining per two field picture;S4, the head shoulder images to extraction are carried out after hsv color spatial transformation, and identification obtains corresponding safety cap wearing state.The present invention is not influenceed by human body attitude, recognition accuracy and efficiency high, and can be very good to evade outdoor optical according to the influence brought, with stronger robustness, can be widely applied in Video Surveillance Industry.
Description
Technical field
The present invention relates to field of video monitoring, more particularly to a kind of safety cap wearing state identification side based on optical flow method
Method and system.
Background technology
Explanation of nouns:
Optical flow method:Defined based on pixel, the collection of all light streams is collectively referred to as optical flow field.It can be obtained by light stream field analysis
To sports ground of the object with respect to observer.The algorithm analyzed in this process turns into optical flow method.
HSV (Hue, Saturation, Value) model:Created according to the intuitive nature of color by A.R.Smith in 1978
A kind of color space built, also referred to as hexagonal pyramid model (Hexcone Model), Model Parameter are respectively tone (H), saturation
Spend (S), brightness (V).
In at the construction field (site), the safety problem of field personnel enjoys people to pay close attention to always, thus inspection personnel whether
Safe wearing cap is very important requirement, is directly connected to the personal safety of field personnel.Existing personal security cap
Recognition methods is less and existing defects, has and is only adapted to upright human body, or the method that face is not blocked, other situations
Artificial observation identification is needed to obtain safety cap wearing state, it is impossible to meet our daily demands.
The content of the invention
In order to solve above-mentioned technical problem, it is an object of the invention to provide the knowledge of the safety cap wearing state based on optical flow method
Other method, it is an object of the invention to provide the safety cap wearing state identifying system based on optical flow method.
The technical solution adopted for the present invention to solve the technical problems is:
Safety cap wearing state recognition methods based on optical flow method, including step:
S1, the monitor video for obtaining job site in real time, and obtain the extraction frame of user preset;
S2, according to default extraction frame, after the picture frame of extract real-time monitor video, Difference Calculation is based on using optical flow method
Carry out movement human Objective extraction;
S3, the movement human target progress head and shoulder detection to extracting acquisition, extract the head shoulder images in obtaining per two field picture;
S4, the head shoulder images to extraction are carried out after hsv color spatial transformation, and identification obtains corresponding safety cap and wears shape
State.
Further, the step S2, including:
S21, according to default extraction frame, extract the picture frame of monitor video successively in real time, obtain sequence of video images;
S22, extracted based on median filtering method from sequence of video images and obtain background image sequence;
S23, sequence of video images and background image sequence be subjected to difference, obtain each difference image;
S24, to difference image carry out binaryzation after, extract obtain movement human target.
Further, the step S24, it is specially:
S24, each pixel value and predetermined threshold value T of difference image sequence be compared, when pixel value is more than default threshold
It is 1 by its binaryzation during value T, and judges point of this as movement human target, otherwise be 0 by its binaryzation, and judges the point
For Background picture point, finally extract and obtain movement human target.
Further, the step S3, including:
S31, the binary image to difference image obtain corresponding vertical projective histogram after projecting;
S32, using the local maximum closest to row coordinate end of vertical projective histogram as head width, and then will
N times of head width is 2.5~3 as head and shoulder height, wherein N span;
S33, interception is carried out to every two field picture according to the head and shoulder of acquisition height obtain head shoulder images.
Further, the step S4, including:
S41, the head shoulder images to extraction carry out hsv color spatial transformation, obtain its h, s, v parameter value;
S42, h, s, v parameter value in the hsv color space obtained according to conversion distribution situation, with reference to default safety cap
Color, identification obtains corresponding safety cap wearing state.
Further, the step S42, it is specially:
The distribution situation of h, s, v parameter value in the hsv color space obtained according to conversion, analysis is obtained in the head shoulder images
Whether the color of head position is fallen into default safety cap color gamut, if, then it represents that the safety cap of corresponding color is worn,
Conversely, representing non-safe wearing cap.
The present invention solves another technical scheme for being used of its technical problem:
Safety cap wearing state identifying system based on optical flow method, including:
Video acquiring module, for obtaining the monitor video of job site in real time, and obtains the extraction frame of user preset;
Extraction module, for according to default extraction frame, after the picture frame of extract real-time monitor video, using optical flow method base
Movement human Objective extraction is carried out in Difference Calculation;
Head and shoulder detection module, for carrying out head and shoulder detection to extracting the movement human target obtained, extracts and obtains per frame figure
Head shoulder images as in;
Identification module, is carried out after hsv color spatial transformation, identification obtains corresponding safety for the head shoulder images to extraction
Cap wearing state.
Further, the extraction module, including:
First extracting sub-module, for according to default extraction frame, extracting the picture frame of monitor video successively in real time, is obtained
Sequence of video images;
Second extracting sub-module, background image sequence is obtained for being extracted based on median filtering method from sequence of video images
Row;
3rd extracting sub-module, for sequence of video images and background image sequence to be carried out into difference, obtains each difference diagram
Picture;
4th extracting sub-module, for being carried out to difference image after binaryzation, extracts and obtains movement human target.
Further, the head and shoulder detection module, including:
First detection sub-module, corresponding upright projection is obtained after being projected for the binary image to difference image
Histogram;
Second detection sub-module, for using the local maximum closest to row coordinate end of vertical projective histogram as
Head width, and then using N times of head width as head and shoulder height, wherein N span is 2.5~3;
3rd detection sub-module, carries out interception to every two field picture for the head and shoulder height according to acquisition and obtains head shoulder images.
Further, the identification module, including:
First identification submodule, carries out hsv color spatial transformation for the head shoulder images to extraction, obtains its h, s, v ginseng
Numerical value;
Second identification submodule, for the distribution situation of h, s, v parameter value in the hsv color space according to conversion acquisition,
With reference to default safety cap color, identification obtains corresponding safety cap wearing state.
The beneficial effects of the invention are as follows:Safety cap wearing state recognition methods based on optical flow method, including step:S1, reality
When obtain the monitor video of job site, and obtain the extraction frame of user preset;S2, according to default extraction frame, extract real-time
After the picture frame of monitor video, Difference Calculation is based on using optical flow method and carries out movement human Objective extraction;S3, to extract obtain
Movement human target carries out head and shoulder detection, extracts the head shoulder images in obtaining per two field picture;S4, the head shoulder images to extraction are carried out
After hsv color spatial transformation, identification obtains corresponding safety cap wearing state.This method is not influenceed by human body attitude, and identification is accurate
True rate and efficiency high, and can be very good to evade outdoor optical according to the influence brought, with stronger robustness.
The present invention another beneficial effect be:The safety cap wearing state identifying system based on optical flow method of the present invention, bag
Include:Video acquiring module, for obtaining the monitor video of job site in real time, and obtains the extraction frame of user preset;Extract mould
Block, for according to default extraction frame, after the picture frame of extract real-time monitor video, being carried out using optical flow method based on Difference Calculation
Movement human Objective extraction;Head and shoulder detection module, for carrying out head and shoulder detection to extracting the movement human target obtained, extraction is obtained
Head shoulder images in obtaining per two field picture;Identification module, carries out after hsv color spatial transformation for the head shoulder images to extraction, knows
Corresponding safety cap wearing state is not obtained.The system is not influenceed by human body attitude, recognition accuracy and efficiency high, and can be with
Evade outdoor optical well according to the influence brought, with stronger robustness.
Brief description of the drawings
The invention will be further described with reference to the accompanying drawings and examples.
Fig. 1 is the flow chart of the safety cap wearing state recognition methods based on optical flow method of the present invention;
Fig. 2 is the vertical projective histogram of head shoulder images in specific embodiment of the invention;
Fig. 3 is the particular flow sheet of the corresponding safety cap wearing state of identification acquisition in specific embodiment of the invention.
Embodiment
Reference picture 1, the invention provides a kind of safety cap wearing state recognition methods based on optical flow method, including step:
S1, the monitor video for obtaining job site in real time, and obtain the extraction frame of user preset;
S2, according to default extraction frame, after the picture frame of extract real-time monitor video, Difference Calculation is based on using optical flow method
Carry out movement human Objective extraction;
S3, the movement human target progress head and shoulder detection to extracting acquisition, extract the head shoulder images in obtaining per two field picture;
S4, the head shoulder images to extraction are carried out after hsv color spatial transformation, and identification obtains corresponding safety cap and wears shape
State.
It is further used as preferred embodiment, the step S2, including:
S21, according to default extraction frame, extract the picture frame of monitor video successively in real time, obtain sequence of video images;
S22, extracted based on median filtering method from sequence of video images and obtain background image sequence;
S23, sequence of video images and background image sequence be subjected to difference, obtain each difference image;
S24, to difference image carry out binaryzation after, extract obtain movement human target.
It is further used as preferred embodiment, the step S24, it is specially:
S24, each pixel value and predetermined threshold value T of difference image sequence be compared, when pixel value is more than default threshold
It is 1 by its binaryzation during value T, and judges point of this as movement human target, otherwise be 0 by its binaryzation, and judges the point
For Background picture point, finally extract and obtain movement human target.
It is further used as preferred embodiment, the step S3, including:
S31, the binary image to difference image obtain corresponding vertical projective histogram after projecting;
S32, using the local maximum closest to row coordinate end of vertical projective histogram as head width, and then will
N times of head width is 2.5~3 as head and shoulder height, wherein N span;
S33, interception is carried out to every two field picture according to the head and shoulder of acquisition height obtain head shoulder images.
It is further used as preferred embodiment, the step S4, including:
S41, the head shoulder images to extraction carry out hsv color spatial transformation, obtain its h, s, v parameter value;
S42, h, s, v parameter value in the hsv color space obtained according to conversion distribution situation, with reference to default safety cap
Color, identification obtains corresponding safety cap wearing state.
It is further used as preferred embodiment, the step S42, it is specially:
The distribution situation of h, s, v parameter value in the hsv color space obtained according to conversion, analysis is obtained in the head shoulder images
Whether the color of head position is fallen into default safety cap color gamut, if, then it represents that the safety cap of corresponding color is worn,
Conversely, representing non-safe wearing cap.
Reference picture 2, present invention also offers the safety cap wearing state identifying system based on optical flow method, including:
Video acquiring module, for obtaining the monitor video of job site in real time, and obtains the extraction frame of user preset;
Extraction module, for according to default extraction frame, after the picture frame of extract real-time monitor video, using optical flow method base
Movement human Objective extraction is carried out in Difference Calculation;
Head and shoulder detection module, for carrying out head and shoulder detection to extracting the movement human target obtained, extracts and obtains per frame figure
Head shoulder images as in;
Identification module, is carried out after hsv color spatial transformation, identification obtains corresponding safety for the head shoulder images to extraction
Cap wearing state.
It is further used as preferred embodiment, the extraction module, including:
First extracting sub-module, for according to default extraction frame, extracting the picture frame of monitor video successively in real time, is obtained
Sequence of video images;
Second extracting sub-module, background image sequence is obtained for being extracted based on median filtering method from sequence of video images
Row;
3rd extracting sub-module, for sequence of video images and background image sequence to be carried out into difference, obtains each difference diagram
Picture;
4th extracting sub-module, for being carried out to difference image after binaryzation, extracts and obtains movement human target.
It is further used as preferred embodiment, the head and shoulder detection module, including:
First detection sub-module, corresponding upright projection is obtained after being projected for the binary image to difference image
Histogram;
Second detection sub-module, for using the local maximum closest to row coordinate end of vertical projective histogram as
Head width, and then using N times of head width as head and shoulder height, wherein N span is 2.5~3;
3rd detection sub-module, carries out interception to every two field picture for the head and shoulder height according to acquisition and obtains head shoulder images.
It is further used as preferred embodiment, the identification module, including:
First identification submodule, carries out hsv color spatial transformation for the head shoulder images to extraction, obtains its h, s, v ginseng
Numerical value;
Second identification submodule, for the distribution situation of h, s, v parameter value in the hsv color space according to conversion acquisition,
With reference to default safety cap color, identification obtains corresponding safety cap wearing state.
The present invention is illustrated below in conjunction with specific embodiment.
Reference picture 1, a kind of safety cap wearing state recognition methods based on optical flow method, including step:
S1, the monitor video for obtaining job site in real time, and obtain the extraction frame of user preset;Enter work in staff
Make to set CCTV camera at the necessary channel in region, monitored in real time, and the default monitoring frame adapted to human body.
S2, according to default extraction frame, after the picture frame of extract real-time monitor video, Difference Calculation is based on using optical flow method
Carry out movement human Objective extraction;Including step S21~S24:
S21, according to default extraction frame, extract the picture frame of monitor video successively in real time, obtain sequence of video images Ik
(x, y) (k=1,2 ..., n);
S22, extracted based on median filtering method from sequence of video images and obtain background image sequence Bk(x,y):Bk(x,y)
=med { IK(x,y)};Med represents that medium filtering is calculated.
S23, sequence of video images and background image sequence be subjected to difference, obtain each difference image dk(x,y):dk(x,y)
=| Ik(x,y)-Bk(x,y)|;
S24, to each difference image dk(x, y) is carried out after binaryzation, is extracted and is obtained movement human target, is specially:Will
Each pixel value of difference image sequence is compared with predetermined threshold value T, when pixel value is more than predetermined threshold value T, by its two-value
1 is turned to, and judges point of this as movement human target, otherwise is 0 by its binaryzation, and judges the point as Background picture point,
Finally extract and obtain movement human target.
Binaryzation result of calculation is as follows:
In above formula, bk(x, y) represents binary value.
S3, the movement human target progress head and shoulder detection to extracting acquisition, extract the head shoulder images in obtaining per two field picture,
Including step S31~S33:
S31, the binary image to difference image obtain corresponding vertical projective histogram after projecting;
S32, using the local maximum closest to row coordinate end of vertical projective histogram as head width, and then will
N times of head width is 2.5~3 as head and shoulder height, wherein N span;
S33, interception is carried out to every two field picture according to the head and shoulder of acquisition height obtain head shoulder images.
In the present embodiment, in the binary image by analyzing difference image in each height pixel shape.It is vertical to throw
Shadow histogram more can accurately react the shape of human body, vertical projective histogram such as Fig. 2 institutes that the present embodiment projection is obtained
Show, in Fig. 2 in the row coordinate representation binary image of abscissa pixel column coordinate, ordinate represents the value of all row in the row
For the number n of 1 pixel.Fig. 2 E points are trough, the maximum of points D for having a part at close crown position A, this
Maximum is the head breadth of human body, according to anthroposomatology general knowledge, and human head and shoulder is highly generally 2.5~3 times of the head breadth, it is possible to
2.5 times of height as head and shoulder model of local value maximum, C points are a 2.5 times of D intersection point, i.e. shoulder position, and B is C points
A subpoint on a horizontal, therefore the present embodiment obtains A points to B points apart from height of the H as head and shoulder model.Only
Want head-and-shoulder area not to be blocked, for the human body of complete human body and top half, use human body vertical projective histogram
The identification of human body and the extraction of head and shoulder model can preferably be realized.
S4, the head shoulder images to extraction are carried out after hsv color spatial transformation, and identification obtains corresponding safety cap and wears shape
State, including step S41 and S42:
S41, the head shoulder images to extraction carry out hsv color spatial transformation, obtain its h, s, v parameter value, transfer process is such as
Under:
If (r, g, b) is the red, green and blue coordinate of color space respectively, their value is the real number between 0 to 1.
For the rgb value (r, g, b) of each pixel, if max is equivalent to value maximum in r, g, b, if min is equivalent in r, g, b most
Small value.The pixel three parameter values of (h, s, v) in HSV are calculated, wherein h ∈ [0,360], s, v ∈ [0,1] calculate public
Formula is:
V=max.
S42, h, s, v parameter value in the hsv color space obtained according to conversion distribution situation, with reference to default safety cap
Color, identification obtains corresponding safety cap wearing state, is specially:H, s, v parameter in the hsv color space obtained according to conversion
Whether the distribution situation of value, the color that analysis obtains head position in the head shoulder images falls into default safety cap color gamut
It is interior, if, then it represents that the safety cap of corresponding color is worn, conversely, representing non-safe wearing cap.Specific deterministic process such as Fig. 3 institutes
Show, first determine whether in head shoulder images, whether the probability of s=0 pixel is more than 0.5, if so, judging v=2 and 3 pixel
Probability whether be more than 0.5, if meet, then it represents that wear white safety cap.Conversely, the probability of s=0 pixel is less than 0.5,
Four kinds of situations:1st, the probability of h=0 pixel is more than 0.5, then it represents that wear red safety cap;2nd, h=2 pixel is general
Rate is more than 0.5, then it represents that wear orange safety cap;3rd, the probability of h=3 pixel is more than 0.5, then it represents that wear yellow safety
Cap;4th, the probability of h=9 pixel is more than 0.5, then it represents that wear blue safety cap.
Therefore, this method, which is extracted, obtains after head shoulder images, can be according to the color gamut of safety cap, to judge to obtain head and shoulder
The safety cap color that movement human target is worn in image, identification obtains the wearing state of safety cap.The present invention is based on light stream
Method carries out safety cap wearing state identification to movement human, is not influenceed by human body attitude, realizes to job site or other need
The region for wanting safe wearing cap to enter accurately is monitored, improves the recognition accuracy and effect of safety cap wearing state
Rate, and this method can be very good to evade outdoor optical according to the influence brought, with stronger robustness.
Safety cap wearing state identifying system based on optical flow method is one-to-one with this method, is repeated no more here.
Above is the preferable implementation to the present invention is illustrated, but the invention is not limited to the implementation
Example, those skilled in the art can also make a variety of equivalent variations or replace on the premise of without prejudice to spirit of the invention
Change, these equivalent modifications or replacement are all contained in the application claim limited range.
Claims (10)
1. the safety cap wearing state recognition methods based on optical flow method, it is characterised in that including step:
S1, the monitor video for obtaining job site in real time, and obtain the extraction frame of user preset;
S2, according to default extraction frame, after the picture frame of extract real-time monitor video, carried out using optical flow method based on Difference Calculation
Movement human Objective extraction;
S3, the movement human target progress head and shoulder detection to extracting acquisition, extract the head shoulder images in obtaining per two field picture;
S4, the head shoulder images to extraction are carried out after hsv color spatial transformation, and identification obtains corresponding safety cap wearing state.
2. the safety cap wearing state recognition methods according to claim 1 based on optical flow method, it is characterised in that the step
Rapid S2, including:
S21, according to default extraction frame, extract the picture frame of monitor video successively in real time, obtain sequence of video images;
S22, extracted based on median filtering method from sequence of video images and obtain background image sequence;
S23, sequence of video images and background image sequence be subjected to difference, obtain each difference image;
S24, to difference image carry out binaryzation after, extract obtain movement human target.
3. the safety cap wearing state recognition methods according to claim 1 based on optical flow method, it is characterised in that the step
Rapid S24, it is specially:
S24, each pixel value and predetermined threshold value T of difference image sequence be compared, when pixel value is more than predetermined threshold value T
When, be 1 by its binaryzation, and judge point of this as movement human target, otherwise be 0 by its binaryzation, and judge the point as
Background picture point, finally extracts and obtains movement human target.
4. the safety cap wearing state recognition methods according to claim 2 based on optical flow method, it is characterised in that the step
Rapid S3, including:
S31, the binary image to difference image obtain corresponding vertical projective histogram after projecting;
S32, using the local maximum closest to row coordinate end of vertical projective histogram as head width, and then by head
N times of width is 2.5~3 as head and shoulder height, wherein N span;
S33, interception is carried out to every two field picture according to the head and shoulder of acquisition height obtain head shoulder images.
5. the safety cap wearing state recognition methods according to claim 2 based on optical flow method, it is characterised in that the step
Rapid S4, including:
S41, the head shoulder images to extraction carry out hsv color spatial transformation, obtain its h, s, v parameter value;
S42, h, s, v parameter value in the hsv color space obtained according to conversion distribution situation, with reference to default safety cap face
Color, identification obtains corresponding safety cap wearing state.
6. the safety cap wearing state recognition methods according to claim 5 based on optical flow method, it is characterised in that the step
Rapid S42, it is specially:
The distribution situation of h, s, v parameter value in the hsv color space obtained according to conversion, analysis obtains head in the head shoulder images
Whether the color of position is fallen into default safety cap color gamut, if, then it represents that the safety cap of corresponding color is worn, instead
It, represents non-safe wearing cap.
7. the safety cap wearing state identifying system based on optical flow method, it is characterised in that including:
Video acquiring module, for obtaining the monitor video of job site in real time, and obtains the extraction frame of user preset;
Extraction module, for according to default extraction frame, after the picture frame of extract real-time monitor video, it to be poor to be based on using optical flow method
Divide to calculate and carry out movement human Objective extraction;
Head and shoulder detection module, for carrying out head and shoulder detection to extracting the movement human target obtained, is extracted in obtaining per two field picture
Head shoulder images;
Identification module, is carried out after hsv color spatial transformation, identification obtains corresponding safety cap and worn for the head shoulder images to extraction
Wear state.
8. the safety cap wearing state identifying system according to claim 7 based on optical flow method, it is characterised in that described to carry
Modulus block, including:
First extracting sub-module, for according to default extraction frame, extracting the picture frame of monitor video successively in real time, obtains video
Image sequence;
Second extracting sub-module, background image sequence is obtained for being extracted based on median filtering method from sequence of video images;
3rd extracting sub-module, for sequence of video images and background image sequence to be carried out into difference, obtains each difference image;
4th extracting sub-module, for being carried out to difference image after binaryzation, extracts and obtains movement human target.
9. the safety cap wearing state identifying system according to claim 8 based on optical flow method, it is characterised in that the head
Shoulder detection module, including:
First detection sub-module, obtains corresponding upright projection Nogata after being projected for the binary image to difference image
Figure;
Second detection sub-module, for regarding the local maximum closest to row coordinate end of vertical projective histogram as head
Width, and then using N times of head width as head and shoulder height, wherein N span is 2.5~3;
3rd detection sub-module, carries out interception to every two field picture for the head and shoulder height according to acquisition and obtains head shoulder images.
10. the safety cap wearing state identifying system according to claim 8 based on optical flow method, it is characterised in that described
Identification module, including:
First identification submodule, carries out hsv color spatial transformation for the head shoulder images to extraction, obtains its h, s, v parameter value;
Second identification submodule, for according to conversion obtain hsv color space h, s, v parameter value distribution situation, with reference to
Default safety cap color, identification obtains corresponding safety cap wearing state.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710188270.3A CN107103617A (en) | 2017-03-27 | 2017-03-27 | The recognition methods of safety cap wearing state and system based on optical flow method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710188270.3A CN107103617A (en) | 2017-03-27 | 2017-03-27 | The recognition methods of safety cap wearing state and system based on optical flow method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107103617A true CN107103617A (en) | 2017-08-29 |
Family
ID=59676175
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710188270.3A Pending CN107103617A (en) | 2017-03-27 | 2017-03-27 | The recognition methods of safety cap wearing state and system based on optical flow method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107103617A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107679524A (en) * | 2017-10-31 | 2018-02-09 | 天津天地伟业信息系统集成有限公司 | A kind of detection method of the safety cap wear condition based on video |
CN109040669A (en) * | 2018-06-28 | 2018-12-18 | 国网山东省电力公司菏泽供电公司 | Intelligent substation video fence method and system |
CN109670441A (en) * | 2018-12-14 | 2019-04-23 | 广东亿迅科技有限公司 | A kind of realization safety cap wearing knows method for distinguishing, system, terminal and computer readable storage medium |
CN109753898A (en) * | 2018-12-21 | 2019-05-14 | 中国三峡建设管理有限公司 | A kind of safety cap recognition methods and device |
CN110210274A (en) * | 2018-02-28 | 2019-09-06 | 杭州海康威视数字技术股份有限公司 | Safety cap detection method, device and computer readable storage medium |
CN111079731A (en) * | 2019-12-03 | 2020-04-28 | 中冶赛迪重庆信息技术有限公司 | Configuration system, method, equipment and medium based on safety helmet identification monitoring system |
CN111414873A (en) * | 2020-03-26 | 2020-07-14 | 广州粤建三和软件股份有限公司 | Alarm prompting method, device and alarm system based on wearing state of safety helmet |
CN112949354A (en) * | 2019-12-10 | 2021-06-11 | 顺丰科技有限公司 | Method and device for detecting wearing of safety helmet, electronic equipment and computer-readable storage medium |
CN114283485A (en) * | 2022-03-04 | 2022-04-05 | 杭州格物智安科技有限公司 | Safety helmet wearing detection method and device, storage medium and safety helmet |
CN115830719A (en) * | 2023-02-16 | 2023-03-21 | 青岛旭华建设集团有限公司 | Construction site dangerous behavior identification method based on image processing |
CN116453100A (en) * | 2023-06-16 | 2023-07-18 | 国家超级计算天津中心 | Method, device, equipment and medium for detecting wearing and taking-off normalization of protective equipment |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102013009A (en) * | 2010-11-15 | 2011-04-13 | 无锡中星微电子有限公司 | Smoke image recognition method and device |
CN104036575A (en) * | 2014-07-01 | 2014-09-10 | 江苏省南京市公路管理处公路科学研究所 | Safety helmet wearing condition monitoring method on construction site |
-
2017
- 2017-03-27 CN CN201710188270.3A patent/CN107103617A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102013009A (en) * | 2010-11-15 | 2011-04-13 | 无锡中星微电子有限公司 | Smoke image recognition method and device |
CN104036575A (en) * | 2014-07-01 | 2014-09-10 | 江苏省南京市公路管理处公路科学研究所 | Safety helmet wearing condition monitoring method on construction site |
Non-Patent Citations (1)
Title |
---|
陈曦: "面向大型工地的视觉监管系统关键技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107679524A (en) * | 2017-10-31 | 2018-02-09 | 天津天地伟业信息系统集成有限公司 | A kind of detection method of the safety cap wear condition based on video |
CN110210274A (en) * | 2018-02-28 | 2019-09-06 | 杭州海康威视数字技术股份有限公司 | Safety cap detection method, device and computer readable storage medium |
CN109040669A (en) * | 2018-06-28 | 2018-12-18 | 国网山东省电力公司菏泽供电公司 | Intelligent substation video fence method and system |
CN109670441B (en) * | 2018-12-14 | 2024-02-06 | 广东亿迅科技有限公司 | Method, system, terminal and computer readable storage medium for realizing wearing recognition of safety helmet |
CN109670441A (en) * | 2018-12-14 | 2019-04-23 | 广东亿迅科技有限公司 | A kind of realization safety cap wearing knows method for distinguishing, system, terminal and computer readable storage medium |
CN109753898A (en) * | 2018-12-21 | 2019-05-14 | 中国三峡建设管理有限公司 | A kind of safety cap recognition methods and device |
CN111079731A (en) * | 2019-12-03 | 2020-04-28 | 中冶赛迪重庆信息技术有限公司 | Configuration system, method, equipment and medium based on safety helmet identification monitoring system |
CN112949354A (en) * | 2019-12-10 | 2021-06-11 | 顺丰科技有限公司 | Method and device for detecting wearing of safety helmet, electronic equipment and computer-readable storage medium |
CN111414873A (en) * | 2020-03-26 | 2020-07-14 | 广州粤建三和软件股份有限公司 | Alarm prompting method, device and alarm system based on wearing state of safety helmet |
CN114283485B (en) * | 2022-03-04 | 2022-10-14 | 杭州格物智安科技有限公司 | Safety helmet wearing detection method and device, storage medium and safety helmet |
CN114283485A (en) * | 2022-03-04 | 2022-04-05 | 杭州格物智安科技有限公司 | Safety helmet wearing detection method and device, storage medium and safety helmet |
CN115830719A (en) * | 2023-02-16 | 2023-03-21 | 青岛旭华建设集团有限公司 | Construction site dangerous behavior identification method based on image processing |
CN115830719B (en) * | 2023-02-16 | 2023-04-28 | 青岛旭华建设集团有限公司 | Building site dangerous behavior identification method based on image processing |
CN116453100A (en) * | 2023-06-16 | 2023-07-18 | 国家超级计算天津中心 | Method, device, equipment and medium for detecting wearing and taking-off normalization of protective equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107103617A (en) | The recognition methods of safety cap wearing state and system based on optical flow method | |
CN106295551B (en) | A kind of personnel safety cap wear condition real-time detection method based on video analysis | |
CN108319934A (en) | Safety cap wear condition detection method based on video stream data | |
CN111191586B (en) | Method and system for inspecting wearing condition of safety helmet of personnel in construction site | |
CN107679524A (en) | A kind of detection method of the safety cap wear condition based on video | |
CN108537256A (en) | A kind of safety cap wears recognition methods and device | |
JP3761059B2 (en) | Method and apparatus for detecting human face and observer tracking display | |
CN104063722B (en) | A kind of detection of fusion HOG human body targets and the safety cap recognition methods of SVM classifier | |
CN104361327B (en) | A kind of pedestrian detection method and system | |
CN108460358A (en) | Safety cap recognition methods based on video stream data | |
CN107085714A (en) | A kind of forest fire detection method based on video | |
CN106548165A (en) | A kind of face identification method of the convolutional neural networks weighted based on image block | |
CN104951773A (en) | Real-time face recognizing and monitoring system | |
CN102622584B (en) | Method for detecting mask faces in video monitor | |
CN104036575A (en) | Safety helmet wearing condition monitoring method on construction site | |
CN106503716A (en) | A kind of safety cap recognition methods that is extracted based on color and contour feature and system | |
CN108710840A (en) | A kind of farmland spray machine device people vision guided navigation path identification method | |
CN112149513A (en) | Industrial manufacturing site safety helmet wearing identification system and method based on deep learning | |
CN106952280B (en) | A kind of spray gun paint amount uniformity detection method based on computer vision | |
CN105844242A (en) | Method for detecting skin color in image | |
CN105160297A (en) | Masked man event automatic detection method based on skin color characteristics | |
CN106096603A (en) | A kind of dynamic flame detection method merging multiple features and device | |
CN112487891B (en) | Visual intelligent dynamic identification model construction method applied to electric power operation site | |
CN104143077B (en) | Pedestrian target search method and system based on image | |
CN109684971A (en) | Magic cube-solving robot algorithm executes method |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170829 |
|
RJ01 | Rejection of invention patent application after publication |