CN109035307A - Setting regions target tracking method and system based on natural light binocular vision - Google Patents
Setting regions target tracking method and system based on natural light binocular vision Download PDFInfo
- Publication number
- CN109035307A CN109035307A CN201810775921.3A CN201810775921A CN109035307A CN 109035307 A CN109035307 A CN 109035307A CN 201810775921 A CN201810775921 A CN 201810775921A CN 109035307 A CN109035307 A CN 109035307A
- Authority
- CN
- China
- Prior art keywords
- interest
- targets
- image
- coordinate
- section
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/292—Multi-camera tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/285—Analysis of motion using a sequence of stereo image pairs
-
- 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
- G06T2207/10021—Stereoscopic video; Stereoscopic image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a kind of setting regions target tracking methods and system based on natural light binocular vision, this method comprises: obtaining the image sequence of two general cameras acquisition；Determine targets of interest constraint condition and depth of field section set by user；Image in image sequence is handled, and using binocular vision track algorithm and the constraint condition of the targets of interest, to treated, image carries out targets of interest extraction operation, determines the coordinate of the targets of interest；The mapping relations of the two-dimensional coordinate of world coordinate system and targets of interest based on double camera and depth of field section set by user, determine image procossing section；The image sequence of two general cameras acquisition in image procossing section is handled, the three-dimensional coordinate of the targets of interest in image procossing section is determined, realizes target tracking.Using method provided by the invention or system, the requirement in secret protection application scenarios can be met according to actual needs, between controlling image treatment region.
Description
Technical field
The present invention relates to technical field of computer vision, in particular to a kind of setting regions based on natural light binocular vision
Target tracking method and system.
Background technique
Human perception external information, 80% or more is obtained by vision.Vision refers not only to the impression to optical signal, also
It include the process of acquisition, processing, transmission, storage and the understanding to visual information.Signal processing theory and computer occur with
Afterwards, it is attempted to obtain ambient image with video camera and is translated into digital signal, realized to visual signal with computer
The overall process of reason, so as to form new one computer vision of subject.Computer vision is exactly to be obtained by imaging sensor
Obtain image or image sequence, and then the hand combined by computer using technologies such as image procossing, pattern-recognition, artificial intelligence
Piecewise analysis understands these images, is described and explains to three-dimensional world.The final goal of computer vision is exactly can part generation
The understanding and understanding to real world are completed for human brain.
In to biological vision systematic research, it is found that nearly all biology with vision there are two eyes, with two
Eyes observe object simultaneously, have the feeling of depth or distance.Therefore in computer vision system, one or more is also commonly used
Same Scene from platform video camera is gone from two or more viewpoints obtains one group of image under different perspectives, then by same
Parallax of one scene point in different images is inferred to the spatial geometric shape of target object and position in scene, this method
Referred to as stereoscopic vision (Stereo Vision), it is an important branch of computer vision and the core of computer vision
Content.
The vision system for having two video cameras is Binocular Stereo Vision System.Binocular stereo vision process and human vision
Three-dimensional perception is closely similar, directly the mode of simulation mankind's eyes processing scenery, easy to be reliable.It is greatly answered due to having
With prospect, all the time attention of the binocular stereo vision by each field scholar.Since the 21th century, with research level
Continue to develop, stereovision technique is applied to the various aspects of social life more and more widely, as industrial products detection with
Measurement, the explanation of the three dimensional analysis of medical image, aerial photograph and satellite photo, three-dimensional map is drawn and mobile robot view
Feel navigation etc..
Moving object detection and tracking technology is another important content of computer vision, and its essence is by camera shooting
The video sequence of machine shooting is analyzed, and is detected the moving target in scene and is carried out feature extraction and tracking, further to mesh
Target kinematic parameter is estimated.Moving object detection and tracking has merged the advanced technology in many fields, including image procossing,
Pattern-recognition, artificial intelligence, automatic control etc., in robot visual guidance, common scene monitoring, military visual guidance, intelligence
Many aspects such as traffic are all widely used.Binocular stereo vision simulates mankind's eyes mechanism, can extract stereo pairs
Between parallax information, this has just restored the three-dimensional information of actual scene to a certain extent, practical when needing to carry out target
Three-dimensional position and depth (i.e. the distance between target and observer) when measuring, binocular stereo vision has its irreplaceable excellent
Point.Therefore, the detection of moving target is studied with tracking under binocular vision environment, it is extensive in accurate tracking, three-dimensional information
The fields such as multiple, target measurement have a very important significance.
Presently mainly input of the image of infrared camera scan as binocular vision system, since infrared imaging only has
There is grayscale information, therefore, has the shortcomings that can not to restore that former phase, equipment cost are high, ambient noise is big, resolution ratio is lower.In addition,
Current binocular vision system is unable to satisfy the requirement in the application scenarios for emphasizing secret protection.
Summary of the invention
The object of the present invention is to provide a kind of setting regions target tracking method and system based on natural light binocular vision,
Between image treatment region can be controlled, meet the requirement in the application scenarios for emphasizing secret protection, while also there is tracking target
The advantages that high-efficient, at low cost and high resolution.
To achieve the above object, the present invention provides following schemes:
A kind of setting regions target tracking method based on natural light binocular vision, the setting regions target tracking method
Include:
Obtain the image sequence of two general cameras acquisition；
Determine the constraint condition of targets of interest；
Image in described image sequence is handled；
Using binocular vision track algorithm and the constraint condition of the targets of interest, to treated, image carries out interest
Objective extraction operation, and determine the coordinate of the targets of interest；The coordinate includes two-dimensional coordinate and three-dimensional coordinate；
Obtain depth of field section set by user；
According to mapping relations of the world coordinate system of double camera and the two-dimensional coordinate of the targets of interest and set by user
Depth of field section determines image procossing section；
The image sequence of two general cameras acquisition in described image processing section is handled, determines image procossing
The three-dimensional coordinate of the targets of interest in section.
Optionally, the image in described image sequence is handled, is specifically included:
Gray processing, binaryzation and the disposal of gentle filter are carried out to the image in described image sequence.
Optionally, described using binocular vision track algorithm and the constraint condition of the targets of interest, to treated
Image carries out targets of interest extraction operation, specifically includes:
Closed regions extraction operation is carried out to treated image, each target object in the image that determines that treated；
Obtain the characteristic information of each target object；The characteristic information includes dimension information, like circularity information, bright
Spend information；
The characteristic information of each target object is compared with the constraint condition of the targets of interest, determines interest
Target.
Optionally, the coordinate of the determination targets of interest, specifically includes:
Camera is demarcated, determines camera parameter；
According to the camera parameter, world coordinate system is established；
According to the world coordinate system and the targets of interest, the coordinate of the targets of interest is determined；The coordinate packet
Include two-dimensional coordinate and three-dimensional coordinate.
Optionally, the characteristic information for obtaining each target object, specifically includes:
The contour line for extracting each target object determines believing like circularity information and size for each target object
Breath；The dimension information includes height, width and area；
Determine the luminance information of each target object.
Optionally, the image sequence to two general cameras acquisition in described image processing section is handled,
The three-dimensional coordinate for determining the targets of interest in image procossing section, specifically includes:
Section is handled according to described image, the image in described image sequence is divided, determines image sequence to be processed
Column return and carry out processing step to the image in described image sequence, and image sequence to be processed is replaced image sequence, until
Determine that the three-dimensional coordinate of the targets of interest in image procossing section terminates.
The present invention also provides a kind of setting regions target tracking system based on natural light binocular vision, the setting regions
Target tracking system includes:
Image sequence obtains module, for obtaining the image sequence of two general cameras acquisition；
Constraint condition determining module, for determining the constraint condition of targets of interest；
Processing module, for handling the image in described image sequence；
Coordinate determining module, for the constraint condition using binocular vision track algorithm and the targets of interest, to place
Image after reason carries out targets of interest extraction operation, and determines the coordinate of the targets of interest；The coordinate includes two-dimensional coordinate
And three-dimensional coordinate；
Depth of field section obtains module, for obtaining depth of field section set by user；
Image procossing section determining module, for being sat according to the world coordinate system of double camera and the two dimension of the targets of interest
Target mapping relations and depth of field section set by user, determine image procossing section；
Targets of interest three-dimensional coordinate determining module in setting range, for common to two in described image processing section
The image sequence of camera acquisition is handled, and determines the three-dimensional coordinate of the targets of interest in image procossing section.
Optionally, the processing module, specifically includes:
Processing unit, for carrying out gray processing, binaryzation and the disposal of gentle filter to the image in described image sequence.
Optionally, the coordinate determining module, specifically includes:
Target object determination unit determines that treated for carrying out Closed regions extraction operation to treated image
Each target object in image；
Characteristic acquisition unit, for obtaining the characteristic information of each target object；The characteristic information includes
Dimension information, like circularity information, luminance information；
Targets of interest determination unit, for by the constraint of the characteristic information of each target object and the targets of interest
Condition is compared, and determines targets of interest；
Camera parameter determination unit determines camera parameter for demarcating to camera；
World coordinate system establishes unit, for establishing world coordinate system according to the camera parameter；
Coordinate determination unit, for determining the targets of interest according to the world coordinate system and the targets of interest
Coordinate；The coordinate includes two-dimensional coordinate and three-dimensional coordinate.
Optionally, the characteristic acquisition unit, specifically includes:
Subelement is determined like circularity information and dimension information, for extracting the contour line of each target object, is determined
Each target object like circularity information and dimension information；The dimension information includes height, width and area；
Luminance information determination unit, for determining the luminance information of each target object.
The specific embodiment provided according to the present invention, the invention discloses following technical effects:
The present invention provides a kind of setting regions target tracking method and system based on natural light binocular vision, this method
It include: the image sequence for obtaining the acquisition of two general cameras；Determine the constraint condition and the depth of field set by user of targets of interest
Section；Image in image sequence is handled, and uses the constraint condition of binocular vision track algorithm and targets of interest,
To treated, image carries out targets of interest extraction operation, determines the coordinate of the targets of interest；It is sat according to the world of double camera
Mapping relations and set by user depth of field section of the mark system with the two-dimensional coordinate of targets of interest, determine image procossing section；To figure
Image sequence as handling two general cameras acquisition in section is handled, and determines the interest in image procossing section
The three-dimensional coordinate of target.Using method provided by the invention, can meet according to actual needs, between controlling image treatment region
Emphasize the requirement in the application scenarios of secret protection.In addition, only handling the image pixel in image procossing section, reduce
Data processing load improves the real-time during tracking.
Detailed description of the invention
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without creative efforts, can also obtain according to these attached drawings
Obtain other attached drawings.
Fig. 1 is that the process of setting regions target tracking method of the embodiment of the present invention based on natural light binocular vision is illustrated
Figure；
Fig. 2 is the embodiment of the present invention using the comparison diagram before and after Anti-interference algorithm；
Fig. 3 is the block diagram of setting regions target tracking system of the embodiment of the present invention based on natural light binocular vision；
Fig. 4 is the structural representation of setting regions target tracking system of the embodiment of the present invention based on natural light binocular vision
Figure.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The object of the present invention is to provide a kind of setting regions target tracking method and system based on natural light binocular vision,
Depth of field parameter can be controlled, the requirement in the application scenarios for emphasizing secret protection is met, while also there is tracking target efficiency
The advantages that high, at low cost and high resolution.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing and specific real
Applying mode, the present invention is described in further detail.
Fig. 1 is that the process of setting regions target tracking method of the embodiment of the present invention based on natural light binocular vision is illustrated
Figure, as shown in Figure 1, the setting regions target tracking method provided in an embodiment of the present invention based on natural light binocular vision specifically wraps
Include following steps.
Step 101: obtaining the image sequence of two general cameras acquisition.
Step 102: determining the constraint condition of targets of interest.
Step 103: the image in described image sequence is handled.
Step 104: using binocular vision track algorithm and the constraint condition of the targets of interest, to treated image
Targets of interest extraction operation is carried out, and determines the coordinate of the targets of interest；The coordinate includes two-dimensional coordinate and three-dimensional coordinate.
Step 105: obtaining depth of field section set by user.
Step 106: according to the mapping relations and use of the world coordinate system of double camera and the two-dimensional coordinate of the targets of interest
The depth of field section of family setting, determines image procossing section.
Step 107: the image sequence of two general cameras acquisition in described image processing section being handled, is determined
The three-dimensional coordinate of the targets of interest in image procossing section.
Step 103 specifically includes:
Gray processing, binaryzation and the disposal of gentle filter are carried out to the image in described image sequence.
The embodiment of the present invention is binocular vision target following to be realized based on natural light, therefore all are close with natural optical wavelength
Light source be likely to cause interference, thus using Anti-interference algorithm handle.
Target object reflects natural light, and camera can be by its image capture.Object has its characteristic feature, such as width, height
Degree, brightness, like circularity etc..In general, needing through some target object of multiple features ability accurate descriptions.The embodiment of the present invention
The Anti-interference algorithm of use is exactly based on series of features to identify targets of interest, filters out non-targeted object.In the specific implementation,
Width, height, the control of area equidimension information are fairly simple.The control of luminance information is mainly used to filter light source interference, real
Processing first is carried out to image now and obtains grayscale image, reduces data processing load, to higher or lower than certain gray scale (threshold value)
Goal-setting " filters out label ".Refer to that target two-dimensional projection tends to circular degree like circularity, mesh is found out by algorithm in realization
Lower limit and the upper limit is arranged like circularity in mark, and the object or light like circularity section, outside section for filtering out interference are defined with this
" filtering out label " will be set in source, finally decide whether to execute filtering according to label weight, this is the main of Anti-interference algorithm
Mentality of designing.Wherein, asking is comparison basis like circularity and mature calculating process, is first found out with function (such as findcontours)
Then objective contour seeks the radius of profile minimum circumscribed circle and minimum inscribed circle respectively, two radiuses are closer, illustrate that target is got over
Close to circle.
The validity of Anti-interference algorithm is shown below by experiment
Experiment condition: under the conditions of natural lighting, select two common cameras as image capture device, the bead that shines is made
To track target.Emphasis detects the stability of Anti-interference algorithm and the stability and real-time of binocular vision track algorithm.
Comparison setting: it compares whether there is or not the detection in the case of Anti-interference algorithm, track situation.
Experimental result: such as Fig. 2, left side is former phase, and centre is the target detection effect without Anti-interference algorithm, and right side is anti-dry
Disturb algorithm effect.As it can be seen that the light source for generating interference to target in image has been filtered, and algorithm has traced into the center of target
Position (black Labeling Coordinate in former phase).
Concrete methods of realizing are as follows: first extract in image that (for computer, profile just represents object for the profile of each target object
Body, while the corresponding number of each profile is to distinguish), and then determine the width of target object, height, like circularity etc.；Figure
As being converted into grayscale image, the brightness of target object is determined.Feature according to the characteristic information of above-mentioned target object, with targets of interest
It is compared, if similarity degree meets the range of setting, then it is assumed that this target object is targets of interest, on the contrary then delete this
A target object.
Therefore step 104 specifically includes:
Closed regions extraction operation is carried out to treated image, each target object in the image that determines that treated.
Obtain the characteristic information of each target object；The characteristic information includes dimension information, like circularity information, bright
Spend information；The contour line for specially extracting each target object, determine each target object like circularity information and
Dimension information；The dimension information includes height, width and area；Determine the luminance information of each target object.
The characteristic information of each target object is compared with the constraint condition of the targets of interest, determines interest
Target.
Camera is demarcated, determines camera parameter.
According to the camera parameter, world coordinate system is established.
According to the world coordinate system and the targets of interest, the coordinate of the targets of interest is determined；The coordinate packet
Include two-dimensional coordinate and three-dimensional coordinate.
Step 106 specifically:
One considerable advantage of binocular vision system is exactly the depth information that can obtain object within sweep of the eye, in order to fill
Divide and utilized, devises depth of field control algolithm.By this depth of field control algolithm can from sterically control image treatment region between,
Object outside depth of field section will not be processed, for there is (such as secret protection of common monitoring system of space protection demand
Problem) application scenarios have definite meaning.
Depth information is had in the three-dimensional coordinate of targets of interest, when depth of field control algolithm executes, in the common of two cameras
In the visual field, certain row distance is separated by each frame image or column distance acquires several targets of interest points, is obtained through three-dimensional localization process
The three-dimensional data of targets of interest point is obtained, wherein including depth information, by this several targets of interest point to drafting based on camera
Then the mapping relations of spatial coordinate and two-dimensional coordinate read the depth of field control interval of user preset, closed according to the mapping of drafting
System's feedback arrives two dimensional image processing unit, is achieved in the image procossing control to setting depth of field section.
Specific mapping relations are depending on application environment and used equipment, because wherein having used camera calibration
Result.So-called " spatial coordinate " is that the coordinate-system based on double camera that camera calibration is established later (is also world coordinates
System), the three-dimensional coordinate in world coordinate system, and one are then calculated according to one group of two-dimensional coordinate of the same point of double camera acquisition
One is corresponding, and then establishes the mapping relations of two-dimensional coordinate and three-dimensional coordinate.
Step 107 specifically includes:
Section is handled according to described image, the image in described image sequence is divided, determines image sequence to be processed
Column, return step 102, and image sequence to be processed is replaced into image sequence, until determining described emerging in image procossing section
The three-dimensional coordinate of interesting target terminates.
Fig. 3 is the block diagram of setting regions target tracking system of the embodiment of the present invention based on natural light binocular vision, such as Fig. 3
Shown, the basic device of the setting regions target tracking system is two general cameras.Camera acquires image in the visual field, in image
Processing stage carries out the processing such as gray processing, binaryzation, smothing filtering, and Anti-interference algorithm extracts contour of object later, according to default
Dimension information, brightness, identify target like features such as circularity and filter non-targeted object, carried out according to features described above in each frame
Pixel matching successively keeps the tracking to target.Camera calibration process obtains the inside and outside parameter of system camera, for establishing the world
Coordinate system calculated and works as in conjunction with the two dimensional image that tracking process obtains according to the three-dimensional system of coordinate of foundation in the three-dimensional localization stage
The three-dimensional coordinate of previous frame target completes positioning.Then it according to depth of field parameter set by user, realizes depth of field control, returns to target
Image processing unit, and then realize the target image processing in setting range, complete target location tracking in setting range.
Fig. 4 is the structural representation of setting regions target tracking system of the embodiment of the present invention based on natural light binocular vision
Figure, as shown in figure 4, a kind of setting regions target tracking system packet based on natural light binocular vision provided in an embodiment of the present invention
It includes:
Image sequence obtains module 100, for obtaining the image sequence of two general cameras acquisition.
Constraint condition determining module 200, for determining the constraint condition of targets of interest.
Processing module 300, for handling the image in described image sequence.
Coordinate determining module 400 is right for the constraint condition using binocular vision track algorithm and the targets of interest
Treated, and image carries out targets of interest extraction operation, and determines the coordinate of the targets of interest；The coordinate includes that two dimension is sat
Mark and three-dimensional coordinate.
Depth of field section obtains module 500, for obtaining depth of field section set by user.
Image procossing section determining module 600, for the world coordinate system and the two of the targets of interest according to double camera
The mapping relations of dimension coordinate and depth of field section set by user, determine image procossing section.
Targets of interest three-dimensional coordinate determining module 700 in setting range, for two in described image processing section
The image sequence of general camera acquisition is handled, and determines the three-dimensional coordinate of the targets of interest in image procossing section.
The processing module 300 specifically includes:
Processing unit, for carrying out gray processing, binaryzation and the disposal of gentle filter to the image in described image sequence.
The coordinate determining module 400 specifically includes:
Target object determination unit determines that treated for carrying out Closed regions extraction operation to treated image
Each target object in image.
Characteristic acquisition unit, for obtaining the characteristic information of each target object；The characteristic information includes
Dimension information, like circularity information, luminance information.
Targets of interest determination unit, for by the constraint of the characteristic information of each target object and the targets of interest
Condition is compared, and determines targets of interest.
Camera parameter determination unit determines camera parameter for demarcating to camera.
World coordinate system establishes unit, for establishing world coordinate system according to the camera parameter.
Coordinate determination unit, for determining the targets of interest according to the world coordinate system and the targets of interest
Coordinate；The coordinate includes two-dimensional coordinate and three-dimensional coordinate.
Wherein, the characteristic acquisition unit, specifically includes:
Subelement is determined like circularity information and dimension information, for extracting the contour line of each target object, is determined
Each target object like circularity information and dimension information；The dimension information includes height, width and area.
Luminance information determination unit, for determining the luminance information of each target object.
The present invention is to be unable to active balance real-time and essence in traditional binocular vision tracking system based on feature training
True property causes not playing effect in the high practical application of some performance requirements, and needs to confine mesh manually in first frame
Mark, also not smart enough and flexible and public video monitoring and secret protection are the backgrounds that current social discusses hot spot in practical application
Under, real-time tracking targets of interest in setting range is realized using Anti-interference algorithm and depth of field control algolithm, protects public affairs well
Many privacies.
Compared with prior art, advantages of the present invention are as follows:
1, infrared camera is compared as acquisition equipment using general camera, can get original and mutually handled and shown, both dropped
Low cost, and meet former mutually this demand of display.
2, using Anti-interference algorithm, annoyance level suffered by the biocular systems based on natural light is reduced to a certain extent, and
The validity of Anti-interference algorithm is shown with experiment.
3, image process range can be controlled according to actual requirement, complexity and the data processing for simplifying feature extraction are negative
Load, improves the real-time of tracking process, for having space protection demand (such as the Privacy Protection of common monitoring system)
Application scenarios have definite meaning.
4, acquisition image can be not only shown but also the image that can show in image procossing section that treated.
Used herein a specific example illustrates the principle and implementation of the invention, and above embodiments are said
It is bright to be merely used to help understand method and its core concept of the invention；At the same time, for those skilled in the art, foundation
Thought of the invention, there will be changes in the specific implementation manner and application range.In conclusion the content of the present specification is not
It is interpreted as limitation of the present invention.
Claims (10)
1. a kind of setting regions target tracking method based on natural light binocular vision, which is characterized in that the setting regions mesh
Marking method for tracing includes:
Obtain the image sequence of two general cameras acquisition；
Determine the constraint condition of targets of interest；
Image in described image sequence is handled；
Using binocular vision track algorithm and the constraint condition of the targets of interest, to treated, image carries out targets of interest
Extraction operation, and determine the coordinate of the targets of interest；The coordinate includes two-dimensional coordinate and three-dimensional coordinate；
Obtain depth of field section set by user；
According to the mapping relations and the depth of field set by user of the world coordinate system of double camera and the two-dimensional coordinate of the targets of interest
Section determines image procossing section；
The image sequence of two general cameras acquisition in described image processing section is handled, determines image procossing section
The three-dimensional coordinate of the interior targets of interest.
2. setting regions target tracking method according to claim 1, which is characterized in that the figure in described image sequence
Picture is handled, and is specifically included:
Gray processing, binaryzation and the disposal of gentle filter are carried out to the image in described image sequence.
3. setting regions target tracking method according to claim 1, which is characterized in that described to be tracked using binocular vision
The constraint condition of algorithm and the targets of interest, to treated, image carries out targets of interest extraction operation, specifically includes:
Closed regions extraction operation is carried out to treated image, each target object in the image that determines that treated；
Obtain the characteristic information of each target object；The characteristic information includes dimension information, like circularity information, brightness letter
Breath；
The characteristic information of each target object is compared with the constraint condition of the targets of interest, determines interest mesh
Mark.
4. setting regions target tracking method according to claim 3, which is characterized in that the determination targets of interest
Coordinate, specifically include
Camera is demarcated, determines camera parameter；
According to the camera parameter, world coordinate system is established；
According to the world coordinate system and the targets of interest, the coordinate of the targets of interest is determined；The coordinate includes two
Tie up coordinate and three-dimensional coordinate.
5. setting regions target tracking method according to claim 3, which is characterized in that described to obtain each target
The characteristic information of object, specifically includes:
The contour line for extracting each target object, determine each target object like circularity information and dimension information；
The dimension information includes height, width and area；
Determine the luminance information of each target object.
6. setting regions target tracking method according to claim 1, which is characterized in that described to described image treatment region
The image sequence of interior two general cameras acquisition is handled, and determines three of the targets of interest in image procossing section
Coordinate is tieed up, is specifically included:
Section is handled according to described image, the image in described image sequence is divided, image sequence to be processed is determined, returns
It returns and processing step is carried out to the image in described image sequence, and image sequence to be processed is replaced into image sequence, until determining
The three-dimensional coordinate of the targets of interest in image procossing section terminates.
7. a kind of setting regions target tracking system based on natural light binocular vision, which is characterized in that the setting regions mesh
Marking tracing system includes:
Image sequence obtains module, for obtaining the image sequence of two general cameras acquisition；
Constraint condition determining module, for determining the constraint condition of targets of interest；
Processing module, for handling the image in described image sequence；
Coordinate determining module, for the constraint condition using binocular vision track algorithm and the targets of interest, after processing
Image carry out targets of interest extraction operation, and determine the coordinate of the targets of interest；The coordinate includes two-dimensional coordinate and three
Tie up coordinate；
Depth of field section obtains module, for obtaining depth of field section set by user；
Image procossing section determining module, for according to the two-dimensional coordinate of the world coordinate system and targets of interest of double camera
Mapping relations and depth of field section set by user, determine image procossing section；
Targets of interest three-dimensional coordinate determining module in setting range, for two general cameras in described image processing section
The image sequence of acquisition is handled, and determines the three-dimensional coordinate of the targets of interest in image procossing section.
8. setting regions target tracking system according to claim 7, which is characterized in that the processing module is specific to wrap
It includes:
Processing unit, for carrying out gray processing, binaryzation and the disposal of gentle filter to the image in described image sequence.
9. setting regions target tracking system according to claim 7, which is characterized in that the coordinate determining module, tool
Body includes:
Target object determination unit, for carrying out Closed regions extraction operation to treated image, the image that determines that treated
In each target object；
Characteristic acquisition unit, for obtaining the characteristic information of each target object；The characteristic information includes size
Information, like circularity information, luminance information；
Targets of interest determination unit, for by the constraint condition of the characteristic information of each target object and the targets of interest
It is compared, determines targets of interest；
Camera parameter determination unit determines camera parameter for demarcating to camera；
World coordinate system establishes unit, for establishing world coordinate system according to the camera parameter；
Coordinate determination unit, for determining the seat of the targets of interest according to the world coordinate system and the targets of interest
Mark；The coordinate includes two-dimensional coordinate and three-dimensional coordinate.
10. setting regions target tracking method according to claim 9, which is characterized in that the characteristic information obtains single
Member specifically includes:
Subelement is determined like circularity information and dimension information, for extracting the contour line of each target object, is determined each
The target object like circularity information and dimension information；The dimension information includes height, width and area；
Luminance information determination unit, for determining the luminance information of each target object.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810775921.3A CN109035307B (en) | 2018-07-16 | 2018-07-16 | Set area target tracking method and system based on natural light binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810775921.3A CN109035307B (en) | 2018-07-16 | 2018-07-16 | Set area target tracking method and system based on natural light binocular vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109035307A true CN109035307A (en) | 2018-12-18 |
| CN109035307B CN109035307B (en) | 2020-09-25 |
Family
ID=64643062
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810775921.3A Active CN109035307B (en) | 2018-07-16 | 2018-07-16 | Set area target tracking method and system based on natural light binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109035307B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110197508A (en) * | 2019-07-10 | 2019-09-03 | 深圳西顺万合科技有限公司 | The method and device of the co-melting vision guide movement of 2D, 3D |
| CN110992393A (en) * | 2019-11-24 | 2020-04-10 | 杭州鼎热科技有限公司 | Target motion tracking method based on vision |
| CN111953933A (en) * | 2020-07-03 | 2020-11-17 | 北京中安安博文化科技有限公司 | Method, device, medium and electronic equipment for determining fire area |
| CN113516709A (en) * | 2021-07-09 | 2021-10-19 | 连云港远洋流体装卸设备有限公司 | Flange positioning method based on binocular vision |
| CN114638750A (en) * | 2022-02-17 | 2022-06-17 | 中山大学 | Parallax enhancement optical telescope and imaging system and method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308018A (en) * | 2008-05-30 | 2008-11-19 | 汤一平 | Stereo vision measuring apparatus based on binocular omnidirectional visual sense sensor |
| CN101883291A (en) * | 2010-06-29 | 2010-11-10 | 上海大学 | Method for drawing viewpoints by reinforcing interested region |
| CN102135236A (en) * | 2011-01-05 | 2011-07-27 | 北京航空航天大学 | Automatic non-destructive testing method for internal wall of binocular vision pipeline |
| CN102699733A (en) * | 2012-06-12 | 2012-10-03 | 大连理工大学 | Method and device for measuring movement locus of automatic tool changing mechanical arm |
| US20170278256A1 (en) * | 2016-03-23 | 2017-09-28 | Akcelita, LLC | System and Method for Tracking and Annotating Multiple Objects in a 3D Model |
| US20180041747A1 (en) * | 2016-08-03 | 2018-02-08 | Samsung Electronics Co., Ltd. | Apparatus and method for processing image pair obtained from stereo camera |
| CN107909604A (en) * | 2017-11-07 | 2018-04-13 | 武汉科技大学 | Dynamic object movement locus recognition methods based on binocular vision |
| CN108234920A (en) * | 2012-12-20 | 2018-06-29 | 微软技术许可有限责任公司 | Video camera with privacy mode |
-
2018
- 2018-07-16 CN CN201810775921.3A patent/CN109035307B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101308018A (en) * | 2008-05-30 | 2008-11-19 | 汤一平 | Stereo vision measuring apparatus based on binocular omnidirectional visual sense sensor |
| CN101883291A (en) * | 2010-06-29 | 2010-11-10 | 上海大学 | Method for drawing viewpoints by reinforcing interested region |
| CN102135236A (en) * | 2011-01-05 | 2011-07-27 | 北京航空航天大学 | Automatic non-destructive testing method for internal wall of binocular vision pipeline |
| CN102699733A (en) * | 2012-06-12 | 2012-10-03 | 大连理工大学 | Method and device for measuring movement locus of automatic tool changing mechanical arm |
| CN108234920A (en) * | 2012-12-20 | 2018-06-29 | 微软技术许可有限责任公司 | Video camera with privacy mode |
| US20170278256A1 (en) * | 2016-03-23 | 2017-09-28 | Akcelita, LLC | System and Method for Tracking and Annotating Multiple Objects in a 3D Model |
| US20180041747A1 (en) * | 2016-08-03 | 2018-02-08 | Samsung Electronics Co., Ltd. | Apparatus and method for processing image pair obtained from stereo camera |
| CN107909604A (en) * | 2017-11-07 | 2018-04-13 | 武汉科技大学 | Dynamic object movement locus recognition methods based on binocular vision |
Non-Patent Citations (3)
| Title |
|---|
| QI GUO等: "Focal Track: Depth and Accommodation with Oscillating Lens Deformation", 《2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION》 * |
| 曾鸿等: "《船舶机电系统虚拟仿真研究》", 30 June 2015 * |
| 王婷婷等: "基于双目视觉的运动目标检测跟踪与定位", 《机械与电子》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110197508A (en) * | 2019-07-10 | 2019-09-03 | 深圳西顺万合科技有限公司 | The method and device of the co-melting vision guide movement of 2D, 3D |
| CN110197508B (en) * | 2019-07-10 | 2024-02-20 | 深圳西顺万合科技有限公司 | 2D and 3D co-fusion vision guiding movement method and device |
| CN110992393A (en) * | 2019-11-24 | 2020-04-10 | 杭州鼎热科技有限公司 | Target motion tracking method based on vision |
| CN111953933A (en) * | 2020-07-03 | 2020-11-17 | 北京中安安博文化科技有限公司 | Method, device, medium and electronic equipment for determining fire area |
| CN111953933B (en) * | 2020-07-03 | 2022-07-05 | 北京中安安博文化科技有限公司 | Method, device, medium and electronic equipment for determining fire area |
| CN113516709A (en) * | 2021-07-09 | 2021-10-19 | 连云港远洋流体装卸设备有限公司 | Flange positioning method based on binocular vision |
| CN113516709B (en) * | 2021-07-09 | 2023-12-29 | 连云港远洋流体装卸设备有限公司 | Flange positioning method based on binocular vision |
| CN114638750A (en) * | 2022-02-17 | 2022-06-17 | 中山大学 | Parallax enhancement optical telescope and imaging system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109035307B (en) | 2020-09-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111062905B (en) | Infrared and visible light fusion method based on saliency map enhancement | |
| CN109035307A (en) | Setting regions target tracking method and system based on natural light binocular vision | |
| CN103868460B (en) | Binocular stereo vision method for automatic measurement based on parallax optimized algorithm | |
| CN112634341B (en) | Method for constructing depth estimation model of multi-vision task cooperation | |
| CN104036488B (en) | Binocular vision-based human body posture and action research method | |
| CN102833487B (en) | Visual computing-based optical field imaging device and method | |
| CN109828658B (en) | Man-machine co-fusion remote situation intelligent sensing system | |
| CN104883556A (en) | Three dimensional display method based on augmented reality and augmented reality glasses | |
| CN106056534A (en) | Obstruction perspective method and device based on smart glasses | |
| CN103796001A (en) | Method and device for synchronously acquiring depth information and color information | |
| CN113362247A (en) | Semantic live-action three-dimensional reconstruction method and system of laser fusion multi-view camera | |
| CN104424640A (en) | Method and device for carrying out blurring processing on images | |
| CN112207821B (en) | Target searching method of visual robot and robot | |
| CN113763231B (en) | Model generation method, image perspective determination method, device, equipment and medium | |
| CN113256699B (en) | Image processing method, image processing device, computer equipment and storage medium | |
| CN109213138A (en) | A kind of barrier-avoiding method, apparatus and system | |
| CN110909571B (en) | High-precision face recognition space positioning method | |
| EP4354401A1 (en) | Method and system of detecting obstacle elements with a visual aid device | |
| CN108564654B (en) | Picture entering mode of three-dimensional large scene | |
| CN117237414A (en) | Grabbing and guiding method and system based on binocular images under mobile robot platform | |
| CN113298177A (en) | Night image coloring method, device, medium, and apparatus | |
| CN113723432B (en) | Intelligent identification and positioning tracking method and system based on deep learning | |
| CN112926498B (en) | Living body detection method and device based on multichannel fusion and depth information local dynamic generation | |
| CN114494582A (en) | Three-dimensional model dynamic updating method based on visual perception | |
| Xu et al. | Real-time panoramic map modeling method based on multisource image fusion and three-dimensional rendering |
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 |