CN109141344A - A kind of method and system based on the accurate ranging of binocular camera - Google Patents
A kind of method and system based on the accurate ranging of binocular camera Download PDFInfo
- Publication number
- CN109141344A CN109141344A CN201810623778.6A CN201810623778A CN109141344A CN 109141344 A CN109141344 A CN 109141344A CN 201810623778 A CN201810623778 A CN 201810623778A CN 109141344 A CN109141344 A CN 109141344A
- Authority
- CN
- China
- Prior art keywords
- data
- camera
- image
- binocular camera
- characteristic
- 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
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C3/00—Measuring distances in line of sight; Optical rangefinders
Abstract
The invention discloses a kind of method and system that distance is accurately measured based on binocular camera.The method includes two main process of calibrating and work.Calibration process calculates one group of characteristic that object embodies on binocular image, and by the storage of the relationship of this group of characteristic and distance into data capsule；Use process, the characteristic embodied on binocular image according to object, extracts actual range from data capsule.The system includes binocular camera, computing module and data capsule.It uses technical solution of the present invention: not needing the range unit additional using laser range finder, ultrasonic sensor etc., advantage of lower cost during the work time；It is not limited by usage scenario；It does not need using complicated algorithm, system resources in computation occupancy is small；The available accurate measurement result of stabilization.
Description
Technical field
The present invention relates to camera ranging field more particularly to a kind of method and system based on binocular camera ranging.
Background technique
Using the method for camera ranging, mainly there are following three types at present:
The first measures camera by installing laser range finder, ultrasonic sensor or other range units additional on camera
To the distance of object, this method is needed all to install additional device additional for every suit equipment, is substantially more expensive.
Second, using monocular cam, pass through some known reference datas and calculate distances, it is known that reference data extremely
Less include one of following data: relative displacement, object between the actual size of object, camera and object with take the photograph
As the angle and camera shooting grease head highness etc. of head mirror head, in this way when must known some data, it is applicable to limit its
Scene.
The third obtains the parameter matrix of camera by stereo calibration using binocular camera, then utilizes calibration knot
Fruit is cooked Stereo matching to binocular image, is generated 3D point cloud by parallax, and then generate depth image, is finally mentioned from depth image
The algorithm for taking camera to need use complicated to the distance of object, this method has large amount of complex in the process and calculates, and by
Stereo calibration and such environmental effects, obtained calculated result are often short of stability.
In the commonly used of artificial intelligence technology, especially on intelligent robot and smart machine, there is an urgent need to one kind
Low manufacture cost, stability and high efficiency and the higher distance measuring method of accuracy and system.
Summary of the invention
In order to solve currently based at high cost present in camera distance measuring method, application scenarios are limited, algorithm is complicated, need
Largely to calculate and the problems such as calculated result is unstable, the invention proposes a kind of method based on binocular camera ranging and
System.The present invention can reuse a set of range unit in mostly set autonomous system, and according to object in binocular camera
In projection quickly obtain and accurately measure as a result, not needing to do camera calibration in the process and using complicated algorithm, will not
Occupy a large amount of system resources in computation.
To achieve the goals above, the invention discloses a kind of methods based on binocular camera ranging, it is characterised in that:
One of camera of binocular camera is defined as benchmark camera, another camera, which is used as, refers to camera,
The image of benchmark camera acquisition is benchmark image, is reference picture with reference to camera acquired image；
It analyzes to binocular camera acquired image, the spy embodied according to projection of the object on two images
It levies data and calculates distance, the characteristic includes that the parallax that object is projected in two images and object are projected in benchmark
Coordinate in the X-axis of image；
Comprising calibration process and use process, calibration process be in order under camera and the existing relative positional relationship of object,
Empirical data is obtained, the foundation of distance is calculated as use process；Projection institute of the use process according to object on two images
The characteristic embodied, extracts range data from empirical data；
More specifically, the calibration process comprises the steps of:
Object is placed on to the front of the binocular camera, guarantees that two cameras can take object, two cameras
An image is acquired respectively；
The parallax and object that calculating object is projected in two images are projected in the coordinate in the X-axis of benchmark image；
Measure the object to the benchmark camera distance；
Record related data, the related data includes: parallax that object is projected in two images, object are projected in benchmark
The distance of coordinate and object in the X-axis of image to camera；
By certain rule, change the relative position of object and binocular camera, repeat the above process, until calibration process knot
Beam.
More specifically, certain rule can be with is defined as: define a unit parallax variable quantity (such as: 1 pixel)；
The X-axis of benchmark image is equally divided into several segments, the vertical line that each cut-point does X-axis is crossed respectively, obtains the collection of a vertical line
It closes；A vertical line in vertical line set is chosen as reference line, the midpoint for allowing object to project is fallen on reference line, passes through change
The variation of one unit occurs for the relative distance of benchmark camera and object, the parallax for being projected in object on two images,
All possibility in specified measurement distance range are traversed, the relation data of one group of parallax and distance on specified reference line is obtained；
Vertical line set is traversed, successively using each vertical line in set as reference line, the operation of previous step is repeated, obtains corresponding benchmark
The relation data of parallax and distance on line.
More specifically, the course of work comprises the steps of:
It extracts object and is projected in the characteristic gone out embodied in the binocular camera image, the characteristic includes:
Object projection parallax in two images and object are projected in the coordinate in the X-axis of benchmark image；
According to the characteristic extracted, or by special algorithm to the data obtained after characteristic working process, from data
Range data is extracted in container.
More specifically, can have a kind of method of simplification: choosing a point in the X-axis of benchmark image (in such as image
Heart point), the vertical line that the point does X-axis is crossed, defining the vertical line is benchmark line, allow object projection fallen on the reference line always,
Because it is definite value that object, which is projected in the coordinate in benchmark image X-axis, this item number can not be recorded in a calibration process
According to, it is only necessary to the parallax and object be projected in two images to object is to camera at the two data
Reason, and be recorded in data capsule, in the course of work, the parallax in two images is projected according to object and is held from data
Range data is extracted in device.
More specifically, after completing the calibration process, if having changed the binocular camera any one camera shooting
The focal length of head, then before calibration result will be no longer accurate, need to recalibrate.
The invention also discloses a kind of systems based on binocular camera ranging, it is characterised in that: includes binocular camera, ranging
Device, computing module and data capsule.
The binocular camera is responsible for acquisition image, and image is passed to the computing module.
The binocular camera can be the camera being integrated into two or more imaging units on one piece of circuit board,
It is also possible to two individual cameras to fix by physics, the camera group for the relative positional relationship for keeping constant the two.
The binocular camera can also include camera holder.
The range unit only uses, the reality for measuring object to the binocular camera in a calibration process
Distance.
The computing module, the image for passing over to the binocular camera are analyzed and processed, and to data
Data in container are operated.In a calibration process, the range data that range finder module passes over is received, is further analyzed
After processing, processing result is recorded in data capsule.During the work time, it is projected in binocular image and is embodied according to object
Characteristic out, extracts range data from data capsule.
The data capsule, for saving the related data in system calibration and the course of work.The related data includes:
The coordinate and object that parallax, the object of object projection in two images are projected in the benchmark image X-axis arrive
The distance of camera.When using a kind of method of simplification described above, the related data includes: object is projected in two width
The distance of parallax and object in image to camera.
Compared with prior art, technical solution of the present invention has at least the following advantages:
1) a set of range unit may be reused, advantage of lower cost；
2) it is not limited by the usage scenario of monocular cam distance measuring method；
3) it does not need largely occupy the computing resource of system using complicated algorithm and a large amount of calculating；
4) available stabilization and accurate measurement result.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of the calibration process of the method for the present invention and system.
Fig. 2 is the structural schematic diagram of the course of work of the method for the present invention and system.
When Fig. 3 fixes in the method for the present invention calibration process, object is projected in the coordinate in the X-axis of benchmark image, no
The schematic diagram of the parallax variation of object projection in two images under same distance.
Fig. 4 is in a kind of calibration process of reduction procedure of the method for the present invention, and object is projected in two width under different distance
The schematic diagram of parallax variation in image.
Specific embodiment
Below with reference to specification and a kind of specific embodiment, property feature is described further for the essence of the present invention.
It is as shown in Figure 1 the structural schematic diagram of the calibration process of presently disclosed method and system, which includes double
Mesh camera, range unit, computing module and data capsule.
It is illustrated in figure 2 the structural schematic diagram of the course of work of presently disclosed method and system, with learning process phase
Than the course of work has lacked range unit, because upon completion of the calibration, in this way it is no longer necessary to which range unit, a set of range unit can
To be reused in multiple independent range-measurement systems.
In the present embodiment, the left side camera for defining binocular camera is benchmark camera, and left side camera collects
Image be benchmark image, right side camera is with reference to camera, and right side camera acquired image is reference picture.
The calibration process of embodiment is as follows:
By binocular camera be fixed on one can calibrated length telescopic rod one end, object is fixed on the other end of telescopic rod
(range unit in the telescopic rod corresponding diagram 1)；In telescopic rod telescopic process, guarantee that the midpoint one of object projection is straight
(i.e. by X-axis) on the perpendicular bisector of X-axis in benchmark image；Before guaranteeing that two cameras can take object
Put, telescopic rod is retracted to it is most short, by under current state parallax and distance the (view in parallax distance relation table is recorded
Gap is from the data capsule in relation table corresponding diagram 1).Telescopic rod is extended, when parallax value of the object in two images occurs one
When the variation of a pixel, a new record is added in parallax distance relation table, until telescopic rod extends to longest, is so far existed
The corresponding distance of different parallaxes will be recorded in parallax distance relation table.The flexible of telescopic rod can be customized according to practical application request
The range units such as radar or laser range finder also can be used instead of telescopic rod in range.
The course of work of embodiment is as follows:
Image is acquired with binocular camera, and image is passed into computing module；Computing module navigates to mesh in benchmark image
Object is marked, the adjustment of the position on benchmark image camera holder is projected according to object, makes benchmark camera face object, i.e.,
On the perpendicular bisector for the X-axis that the central point for allowing object to project falls in benchmark image or fall near X-axis perpendicular bisector
In one fixed area；It calculates object under current state and is projected in the parallax in two images, mentioned from parallax distance relation table
Take range data corresponding to the parallax.
In a calibration process, it can be obtained according to capacity of equipment and actual needs and record empirical value, to reach accurate reality
Purpose.
Finally it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention rather than protects to the present invention
The limitation of range, although the invention is described in detail with reference to the preferred embodiments, those skilled in the art should be managed
Solution, modification or equivalent replacement of the technical solution of the present invention are made, without departing from the objective and range of technical solution of the present invention,
It is intended to be within the scope of the claims of the invention.
Claims (10)
1. a kind of method based on binocular camera ranging, it is characterised in that:
Include calibration process and use process；
Empirical data is obtained according to actual range during calibration, and is stored data into data capsule, what is used
In the process, the characteristic formed on two images by object inquires from data capsule according to obtained feature and obtains phase
The accurate distance value answered；
The characteristic includes that object is projected in the parallax in two images.
2. calibration process according to claim 1, which is characterized in that comprise the steps of:
Object is placed on to the front of the binocular camera, guarantees that two cameras can take object, two cameras
An image is acquired respectively；
It calculates object and projects the characteristic embodied in the picture；
Measure the object to the benchmark camera distance；
Related data is saved in data capsule；
By certain rule, change the relative position of object and binocular camera, repeat the above process, until calibration process knot
Beam.
3. the course of work according to claim 1, which is characterized in that comprise the steps of:
It extracts object and is projected in the characteristic gone out embodied in binocular image；
According to the characteristic extracted, or by special algorithm to the data obtained after characteristic working process, from data
Range data is extracted in container.
4. according to claim 1 or claim 2 or characteristic as claimed in claim 3, it is characterised in that: also include target
Object is projected in the coordinate in the X-axis of benchmark image.
5. certain rule according to claim 2, which is characterized in that can be with is defined as:
Define a unit parallax variable quantity (such as: 1 pixel)；
The X-axis of benchmark image is equally divided into several segments, the vertical line that each cut-point does X-axis is crossed respectively, obtains a vertical line
Set；
A vertical line in vertical line set is chosen as reference line, the midpoint for allowing object to project is fallen on reference line, by changing
The change of one unit occurs for the relative distance of variable reference camera and object, the parallax for being projected in object on two images
Change, traverse all possibility in specified measurement distance range, obtains the relationship number of one group of parallax and distance on specified reference line
According to；
Vertical line set is traversed, successively using each vertical line in set as reference line, repeats the operation of previous step, is obtained corresponding
The relation data of parallax and distance on reference line.
6. a kind of system based on binocular camera ranging, it is characterised in that:
Include binocular camera, range unit, computing module and data capsule；
The binocular camera is responsible for acquisition image, and image is passed to the computing module；
The range unit, the actual range for measuring object to the binocular camera；
The computing module, the image for passing over to the binocular camera are analyzed and processed, and to data capsule
In data operated；
In a calibration process, the range data that range finder module passes over is received, after further analysis processing, by processing result
It is recorded in data capsule；
During the work time, it is projected in the characteristic embodied in binocular image according to object, is extracted from data capsule
Range data；
The data capsule, for saving the related data in system calibration and the course of work.
7. binocular camera according to claim 6, which is characterized in that also include: camera holder.
8. range unit according to claim 6, it is characterised in that: only use in a calibration process.
9. the related data that data capsule is saved according to claim 6, it is characterised in that: be projected in two comprising object
The distance of parallax, object in width image to camera.
10. the related data that data capsule is saved according to claim 6, it is characterised in that: also projected comprising object
Coordinate in the benchmark image X-axis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810623778.6A CN109141344A (en) | 2018-06-15 | 2018-06-15 | A kind of method and system based on the accurate ranging of binocular camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810623778.6A CN109141344A (en) | 2018-06-15 | 2018-06-15 | A kind of method and system based on the accurate ranging of binocular camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109141344A true CN109141344A (en) | 2019-01-04 |
Family
ID=64802056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810623778.6A Pending CN109141344A (en) | 2018-06-15 | 2018-06-15 | A kind of method and system based on the accurate ranging of binocular camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109141344A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147759A (en) * | 2020-02-19 | 2020-05-12 | 北京中安吉泰科技有限公司 | Binocular range finding cloud platform device |
| CN111551920A (en) * | 2020-04-16 | 2020-08-18 | 重庆大学 | Three-dimensional target real-time measurement system and method based on target detection and binocular matching |
| CN112489104A (en) * | 2020-12-03 | 2021-03-12 | 海宁奕斯伟集成电路设计有限公司 | Distance measurement method and device, electronic equipment and readable storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6993179B1 (en) * | 2000-08-07 | 2006-01-31 | Koninklijke Philips Electronics N.V. | Strapdown system for three-dimensional reconstruction |
| CN101706957A (en) * | 2009-10-30 | 2010-05-12 | 无锡景象数字技术有限公司 | Self-calibration method for binocular stereo vision device |
| CN102867304A (en) * | 2012-09-04 | 2013-01-09 | 南京航空航天大学 | Method for establishing relation between scene stereoscopic depth and vision difference in binocular stereoscopic vision system |
| CN105091849A (en) * | 2014-05-05 | 2015-11-25 | 南京理工大学 | Optical axis nonlinear binocular range finding method |
| CN105277169A (en) * | 2015-09-25 | 2016-01-27 | 安霸半导体技术（上海）有限公司 | Image segmentation-based binocular range finding method |
| CN105716568A (en) * | 2016-01-28 | 2016-06-29 | 武汉光庭信息技术股份有限公司 | Binocular camera ranging method in automatic pilot system |
| CN106504248A (en) * | 2016-12-06 | 2017-03-15 | 成都通甲优博科技有限责任公司 | Vehicle damage method of discrimination based on computer vision |
| CN103606149B (en) * | 2013-11-14 | 2017-04-19 | 深圳先进技术研究院 | Method and apparatus for calibration of binocular camera and binocular camera |
| CN107093195A (en) * | 2017-03-10 | 2017-08-25 | 西北工业大学 | A kind of locating mark points method that laser ranging is combined with binocular camera |
| WO2017205102A1 (en) * | 2016-05-23 | 2017-11-30 | Microsoft Technology Licensing, Llc | Imaging system comprising real-time image registration |
-
2018
- 2018-06-15 CN CN201810623778.6A patent/CN109141344A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6993179B1 (en) * | 2000-08-07 | 2006-01-31 | Koninklijke Philips Electronics N.V. | Strapdown system for three-dimensional reconstruction |
| CN101706957A (en) * | 2009-10-30 | 2010-05-12 | 无锡景象数字技术有限公司 | Self-calibration method for binocular stereo vision device |
| CN102867304A (en) * | 2012-09-04 | 2013-01-09 | 南京航空航天大学 | Method for establishing relation between scene stereoscopic depth and vision difference in binocular stereoscopic vision system |
| CN103606149B (en) * | 2013-11-14 | 2017-04-19 | 深圳先进技术研究院 | Method and apparatus for calibration of binocular camera and binocular camera |
| CN105091849A (en) * | 2014-05-05 | 2015-11-25 | 南京理工大学 | Optical axis nonlinear binocular range finding method |
| CN105277169A (en) * | 2015-09-25 | 2016-01-27 | 安霸半导体技术（上海）有限公司 | Image segmentation-based binocular range finding method |
| CN105716568A (en) * | 2016-01-28 | 2016-06-29 | 武汉光庭信息技术股份有限公司 | Binocular camera ranging method in automatic pilot system |
| WO2017205102A1 (en) * | 2016-05-23 | 2017-11-30 | Microsoft Technology Licensing, Llc | Imaging system comprising real-time image registration |
| CN106504248A (en) * | 2016-12-06 | 2017-03-15 | 成都通甲优博科技有限责任公司 | Vehicle damage method of discrimination based on computer vision |
| CN107093195A (en) * | 2017-03-10 | 2017-08-25 | 西北工业大学 | A kind of locating mark points method that laser ranging is combined with binocular camera |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147759A (en) * | 2020-02-19 | 2020-05-12 | 北京中安吉泰科技有限公司 | Binocular range finding cloud platform device |
| CN111147759B (en) * | 2020-02-19 | 2021-07-13 | 北京中安吉泰科技有限公司 | Binocular range finding cloud platform device |
| CN111551920A (en) * | 2020-04-16 | 2020-08-18 | 重庆大学 | Three-dimensional target real-time measurement system and method based on target detection and binocular matching |
| CN112489104A (en) * | 2020-12-03 | 2021-03-12 | 海宁奕斯伟集成电路设计有限公司 | Distance measurement method and device, electronic equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111354042B (en) | Feature extraction method and device of robot visual image, robot and medium | |
| CN105627932B (en) | A kind of distance measuring method and device based on binocular vision | |
| CN112927280B (en) | Method and device for acquiring depth image and monocular speckle structured light system | |
| US11132810B2 (en) | Three-dimensional measurement apparatus | |
| CN109360239B (en) | Obstacle detection method, obstacle detection device, computer device, and storage medium | |
| CN105451012B (en) | 3-D imaging system and three-D imaging method | |
| DE102015015194A1 (en) | Image processing apparatus and method and program | |
| CN109141344A (en) | A kind of method and system based on the accurate ranging of binocular camera | |
| CN111862180B (en) | Camera set pose acquisition method and device, storage medium and electronic equipment | |
| CN105378794A (en) | 3d recording device, method for producing 3d image, and method for setting up 3d recording device | |
| JP2019190974A (en) | Calibration device, calibration method and program | |
| JPWO2011125937A1 (en) | Calibration data selection device, selection method, selection program, and three-dimensional position measurement device | |
| CN111951326A (en) | Target object skeleton key point positioning method and device based on multiple camera devices | |
| CA3233222A1 (en) | Method, apparatus and device for photogrammetry, and storage medium | |
| US20160232654A1 (en) | Information processing apparatus, control method thereof, and computer readable storage medium | |
| CN113014816A (en) | Method and device for determining trigger point of flying shooting | |
| CN111798507A (en) | Power transmission line safety distance measuring method, computer equipment and storage medium | |
| US20170099477A1 (en) | Method and system for recalibrating sensing devices without familiar targets | |
| CN110260801A (en) | Method and apparatus for measuring volume of material | |
| CN112771575A (en) | Distance determination method, movable platform and computer readable storage medium | |
| JP2019109200A (en) | Data generation device for calibration, data generation method for calibration, calibration system, and control program | |
| CN111583338B (en) | Positioning method and device for unmanned equipment, medium and unmanned equipment | |
| CN110322518B (en) | Evaluation method, evaluation system and test equipment of stereo matching algorithm | |
| JP2012160063A (en) | Sphere detection method | |
| CN115018922A (en) | Distortion parameter calibration method, electronic device and computer readable storage medium |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190104 |