CN109299665A - A kind of humanoid profile based on LSD algorithm describes method - Google Patents
A kind of humanoid profile based on LSD algorithm describes method Download PDFInfo
- Publication number
- CN109299665A CN109299665A CN201810998054.XA CN201810998054A CN109299665A CN 109299665 A CN109299665 A CN 109299665A CN 201810998054 A CN201810998054 A CN 201810998054A CN 109299665 A CN109299665 A CN 109299665A
- Authority
- CN
- China
- Prior art keywords
- point
- grid
- principal direction
- gradient
- motor point
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of humanoid profiles based on LSD algorithm to describe method, characterized by comprising the following steps: line segment information and output data in input data, the gradient information for calculating the institute motor point, the principal direction point for gridding processing, traversing the motor point, updating the grid, the principal direction for traversing the motor point again, updating the grid using LSD algorithm, all grids of calculating;Time-consuming humanoid judgement that is low and can carrying out each visual angle of the invention is applicable to shops's monitoring shooting visual angle multiplicity, infrared mode low contrast, the feature of scene complexity.
Description
Technical field
This method designed image process field, belongs to the image characteristic extracting method of art of image analysis.
Background technique
Currently, most important humanoid feature extracting method is characterized in based on HOG feature and CNN depth convolution, both
Character description method dimension is high, needs that machine learning algorithm is combined to determine humanoid, and the time-consuming height of one side target detection is not
Meet the requirement applied in real time, is on the other hand not applied for security protection multi-angle of view using the decision model that machine learning algorithm obtains
Application scenarios;Actual shop monitoring has scene complicated, and shooting visual angle is various, and the low spy of picture quality under infrared mode
Point, original HOG feature and CNN depth convolution feature do not adapt to the application demand of actual store monitoring.
Summary of the invention
In view of this, the present invention provides a kind of humanoid profiles based on LSD algorithm to describe method, which is characterized in that packet
Include following steps:
(1) input data, data are the movement point set that moving object detection arrives, and are executed step (2);
(2) gradient information for calculating institute motor point calculates the gradient value and gradient direction in motor point using grayscale image, and
Gradient direction is normalized into [0,180] range, is executed step (3);
(3) gridding is handled, and is carried out 8*8 gridding spatially to movement point set and is divided, and initializes in grid
Main side's point is sky, and principal direction is -1, is executed step (4);
(4) coverage motion point otherwise obtains current kinetic point P and executes step (5) if traversal terminates to execute step (6);
(5) the principal direction point for updating grid, calculates the gridding coordinate (Yp, Xp) of current kinetic point P, if grid (Yp,
Xp the principal direction in) is that the gradient value Grad (P0) of -1 or existing principal direction point P0 is less than the gradient value of current kinetic point P
Grad (P) then updates the gradient direction that the principal direction in grid (Yp, Xp) is motor point P, and the principal direction point for updating grid is P,
Continue to execute step (4);
(6) coverage motion point otherwise obtains current kinetic point P and executes step if traversal terminates to execute step (8) again
(7) step (8) are executed;
(7) principal direction that grid is updated using LSD algorithm, calculates the gridding coordinate (Yp, Xp) of current kinetic point P, such as
Principal direction in fruit grid (Yp, Xp) is differed with the gradient direction of current kinetic point P less than 30 degree, then using in LSD algorithm
Direction merging method update the principal direction in grid (Yp, Xp), and by current kinetic point P accumulation to grid (Yp, Xp) main side
Into point, step (6) are continued to execute;
(8) the line segment information in all grids is calculated, initialization humanoid profile information chained list L is sky, all grids are traversed,
Local line's segment information in grid, the direction including line segment in grid, gradient magnitude and center location information are calculated, if net
Principal direction point number in lattice is greater than 3, then it is assumed that the local line segment in grid is effective, and effective local line's segment information is inserted into
In profile information chained list L, after traversal, execute step (9);
(9) output data exports the humanoid profile information chained list L being made of local lines.
Beneficial achievement of the invention are as follows: this method is extracted humanoid wheel by LSD algorithm on the basis of the motor point
On the one hand wide feature reduces interference of the motor point described in redundancy to humanoid profile, provides algorithm robustness, on the other hand will
Humanoid profile is described in the form of local line segment, and local line's segment information includes local direction, and strength information facilitates and uses mode
The mode of judgement carries out humanoid judgement, so that humanoid determine application scenarios that are high-efficient, and being adapted to market monitoring multi-angle of view, tool
There are vast market prospect and application value.
Specific embodiment
In order to which technical problems, technical solutions and advantages to be solved are more clearly understood, tie below
Embodiment is closed, the present invention will be described in detail.It should be noted that specific embodiment described herein is only to explain
The present invention is not intended to limit the present invention, and the product for being able to achieve said function belongs to equivalent replacement and improvement, is all contained in this hair
Within bright protection scope.The specific method is as follows:
Embodiment 1: in view of this, the present invention provides a kind of humanoid profiles based on LSD algorithm to describe method, feature
It is, includes the following steps:
(1) input data, data are the movement point set that moving object detection arrives, and are executed step (2);
(2) gradient information for calculating institute motor point calculates the gradient value and gradient direction in motor point using grayscale image, and
Gradient direction is normalized into [0,180] range, is executed step (3);
(3) gridding is handled, and is carried out 8*8 gridding spatially to movement point set and is divided, and initializes in grid
Main side's point is sky, and principal direction is -1, is executed step (4);
(4) coverage motion point otherwise obtains current kinetic point P and executes step (5) if traversal terminates to execute step (6);
(5) the principal direction point for updating grid, calculates the gridding coordinate (Yp, Xp) of current kinetic point P, if grid (Yp,
Xp the principal direction in) is that the gradient value Grad (P0) of -1 or existing principal direction point P0 is less than the gradient value of current kinetic point P
Grad (P) then updates the gradient direction that the principal direction in grid (Yp, Xp) is motor point P, and the principal direction point for updating grid is P,
Continue to execute step (4);
(6) coverage motion point otherwise obtains current kinetic point P and executes step if traversal terminates to execute step (8) again
(7) step (8) are executed;
(7) principal direction that grid is updated using LSD algorithm, calculates the gridding coordinate (Yp, Xp) of current kinetic point P, such as
Principal direction in fruit grid (Yp, Xp) is differed with the gradient direction of current kinetic point P less than 30 degree, then using in LSD algorithm
Direction merging method update the principal direction in grid (Yp, Xp), and by current kinetic point P accumulation to grid (Yp, Xp) main side
Into point, step (6) are continued to execute;
(8) the line segment information in all grids is calculated, initialization humanoid profile information chained list L is sky, all grids are traversed,
Local line's segment information in grid, the direction including line segment in grid, gradient magnitude and center location information are calculated, if net
Principal direction point number in lattice is greater than 3, then it is assumed that the local line segment in grid is effective, and effective local line's segment information is inserted into
In profile information chained list L, after traversal, execute step (9);
(9) output data exports the humanoid profile information chained list L being made of local lines.
Claims (1)
1. the present invention relates to a kind of humanoid profiles based on LSD algorithm to describe method, which comprises the steps of:
(1) input data, the data are the movement point set that moving object detection arrives, and are executed step (2);
(2) calculate the motor point gradient information, gradient value and the gradient side in the motor point are calculated using grayscale image
To, and gradient direction is normalized into [0,180] range, it executes step (3);
(3) gridding is handled, and is carried out 8*8 gridding spatially to the movement point set and is divided, and initializes the grid
Interior main side's point is sky, and principal direction is -1, is executed step (4);
(4) motor point is traversed, if traversal terminates to execute step (6), presently described motor point P is otherwise obtained and executes step
(5);
(5) the principal direction point for updating the grid, calculates the gridding coordinate (Yp, Xp) of presently described motor point P, if described
The gradient value Grad (P0) that principal direction in grid (Yp, Xp) is -1 or existing principal direction point P0 is less than presently described movement
The gradient value Grad (P) of point P then updates the gradient direction that the principal direction in the grid (Yp, Xp) is the motor point P, more
The principal direction point of the new grid is P, continues to execute step (4);
(6) motor point is traversed again, if traversal terminates to execute step (8), is otherwise obtained presently described motor point P and is executed
Step (7) executes step (8);
(7) principal direction that the grid is updated using LSD algorithm, calculates the gridding coordinate of presently described motor point P
(Yp, Xp), if principal direction in the grid (Yp, Xp) is differed with the gradient direction of presently described motor point P less than 30
Degree then updates the principal direction in the grid (Yp, Xp) using the direction merging method in the LSD algorithm, and by current institute
It states in motor point P accumulation to the grid (Yp, Xp) principal direction point, continues to execute step (6);
(8) the line segment information in all grids is calculated, initialization humanoid profile information chained list L is sky, all grids are traversed,
Calculate local line's segment information in grid, the direction including the line segment in the grid, gradient magnitude and centre bit confidence
Breath, if the principal direction point number in grid is greater than 3, then it is assumed that the local line segment in the grid is effective, and will be described
Effective local line segment information is inserted into the profile information chained list L, after traversal, is executed step (9);
(9) output data, the humanoid profile information chained list L that output is made of the local lines.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810998054.XA CN109299665B (en) | 2018-08-29 | 2018-08-29 | Human-shaped contour description method based on LSD algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810998054.XA CN109299665B (en) | 2018-08-29 | 2018-08-29 | Human-shaped contour description method based on LSD algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109299665A true CN109299665A (en) | 2019-02-01 |
CN109299665B CN109299665B (en) | 2023-04-14 |
Family
ID=65165954
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810998054.XA Active CN109299665B (en) | 2018-08-29 | 2018-08-29 | Human-shaped contour description method based on LSD algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109299665B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10143654A (en) * | 1996-11-08 | 1998-05-29 | Sony Corp | Outline extracting device and outline extracting method |
CN105550678A (en) * | 2016-02-03 | 2016-05-04 | 武汉大学 | Human body motion feature extraction method based on global remarkable edge area |
CN107194940A (en) * | 2017-05-23 | 2017-09-22 | 北京计算机技术及应用研究所 | A kind of coloured image contour extraction method based on color space and line segment |
-
2018
- 2018-08-29 CN CN201810998054.XA patent/CN109299665B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10143654A (en) * | 1996-11-08 | 1998-05-29 | Sony Corp | Outline extracting device and outline extracting method |
CN105550678A (en) * | 2016-02-03 | 2016-05-04 | 武汉大学 | Human body motion feature extraction method based on global remarkable edge area |
CN107194940A (en) * | 2017-05-23 | 2017-09-22 | 北京计算机技术及应用研究所 | A kind of coloured image contour extraction method based on color space and line segment |
Non-Patent Citations (1)
Title |
---|
睢丹等: "多轮廓三维立体视景图像的投影检测算法", 《吉林大学学报(理学版)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN109299665B (en) | 2023-04-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106407928B (en) | Transformer composite insulator casing monitoring method and system based on raindrop identification | |
CN103559703A (en) | Crane barrier monitoring and prewarning method and system based on binocular vision | |
CN108090572B (en) | Control method of offshore wind farm augmented reality system | |
CN103729854B (en) | A kind of method for detecting infrared puniness target based on tensor model | |
CN104268519B (en) | Image recognition terminal and its recognition methods based on pattern match | |
CN106707296A (en) | Dual-aperture photoelectric imaging system-based unmanned aerial vehicle detection and recognition method | |
CN110349207A (en) | A kind of vision positioning method under complex environment | |
CN104091348A (en) | Multi-target tracking method integrating obvious characteristics and block division templates | |
CN105184812A (en) | Target tracking-based pedestrian loitering detection algorithm | |
CN104794435A (en) | Unmanned aerial vehicle to ground moving target detection method based on vision | |
Kothiya et al. | A review on real time object tracking in video sequences | |
CN110133443B (en) | Power transmission line component detection method, system and device based on parallel vision | |
Pei et al. | Effective image registration based on improved harris corner detection | |
CN107808524A (en) | A kind of intersection vehicle checking method based on unmanned plane | |
CN107423709A (en) | A kind of object detection method for merging visible ray and far infrared | |
CN103985141B (en) | Method for tracking target based on hsv color covariance feature | |
Tashlinskii et al. | Pixel-by-pixel estimation of scene motion in video | |
CN116429082A (en) | Visual SLAM method based on ST-ORB feature extraction | |
Chen et al. | A bridge crack image detection and classification method based on climbing robot | |
CN109299665A (en) | A kind of humanoid profile based on LSD algorithm describes method | |
CN110377033B (en) | RGBD information-based small football robot identification and tracking grabbing method | |
CN104240268B (en) | A kind of pedestrian tracting method based on manifold learning and rarefaction representation | |
CN101540833B (en) | Anti-interference real-time tracking method for profile of object | |
CN110660081B (en) | Target tracking method based on self-adaptive feature selection | |
CN112818881B (en) | Human behavior recognition 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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A Method for Describing Human Figure Contours Based on LSD Algorithm Effective date of registration: 20230701 Granted publication date: 20230414 Pledgee: Bank of Beijing Limited by Share Ltd. Shanghai branch Pledgor: SHANGHAI ULUCU ELECTRONIC TECHNOLOGY CO.,LTD. Registration number: Y2023980046838 |
|
PE01 | Entry into force of the registration of the contract for pledge of patent right |