CN108548482A - A kind of pattern recognition device and method based on monocular vision and laser sensing - Google Patents
A kind of pattern recognition device and method based on monocular vision and laser sensing Download PDFInfo
- Publication number
- CN108548482A CN108548482A CN201810171987.1A CN201810171987A CN108548482A CN 108548482 A CN108548482 A CN 108548482A CN 201810171987 A CN201810171987 A CN 201810171987A CN 108548482 A CN108548482 A CN 108548482A
- Authority
- CN
- China
- Prior art keywords
- laser
- ring
- pen
- monocular vision
- ccd camera
- 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
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/002—Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- 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
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The present invention relates to a kind of pattern recognition device based on monocular vision and laser sensing, including holder, LED light, CCD camera, computer, ring laser sensor；Holder is erected at the surface of object under test, and the LED light that is illuminated to object, CCD camera, the ring laser sensor shot to object are installed on holder；Ring laser sensor includes that can swing the laser pen that can be also slid along annulus, to which laser pen forms at least three laser points in body surface；Computer obtains position and the pivot angle information of laser pen, and the location information of laser point is obtained by CCD camera, finally obtains the spatial positional information of object.Further relate to a kind of image-recognizing method based on monocular vision and laser sensing.Method of the present invention by calculating normal vector, easy to operate, identification is accurate and recognition speed is fast, and work efficiency is high, significantly reduces the working strength of worker, belongs to field of machine vision.
Description
Technical field
The present invention relates to the analyses that field of machine vision more particularly to a kind of laser sensing and camera identify image, pass through
Algorithm calculates the identification device and method of object plane normal vector.
Background technology
With the development of industrial automation technology, for producing the detection with workpiece on logistics transmission assembly line line, due to
The speed of service is fast, and workpiece type is more, and distance change is big etc., and current workpiece identification technology seems helpless.Especially for
Background complex environment, complicated partitioning algorithm between the tiny workpiece of diversity and assembly line of color target, although can
Accurately identification target, but the speed of service is slow, causes robot manipulating task inefficient, directly affects the production effect of assembly line
Rate.
In order to improve the discrimination of target workpiece on assembly line line, simple operations, the algorithm that real-time is high, the speed of service is fast
The hot spot explored as target image.
Invention content
For the technical problems in the prior art, the purpose of the present invention is：There is provided a kind of installation it is simple, quickly to know
Not and capture the pattern recognition device based on monocular vision and laser sensing that object provides support.
Another object of the present invention is：A kind of image-recognizing method based on monocular vision and laser sensing is provided, is dropped
The R＆D costs of vision module on low assembly line, the device installation that the present invention designs is simple, and the principle of calculating is easy-to-understand, laser
After pen measures workpiece depth information and pivot angle information, data are substituted into mathematics geometric knowledge model, can quickly calculate object
Space plane normal vector.
In order to achieve the above object, the present invention adopts the following technical scheme that：
A kind of pattern recognition device based on monocular vision and laser sensing, including holder, LED light, CCD camera, meter
Calculation machine, ring laser sensor；Holder is erected at the surface of object under test, LED light that object is illuminated, to object into
CCD camera, the ring laser sensor of row shooting are installed on holder；Ring laser sensor includes that can swing also can
Along the laser pen that annulus slides, to which laser pen forms at least three laser points in body surface；Computer obtains laser pen
Position and pivot angle information, by CCD camera obtain laser point location information, finally obtain the space bit confidence of object
Breath.
As a preferred embodiment, ring laser sensor includes ring-shaped guide rail, movable plate, chaining part, laser pen, sliding driving
Component, wobble drive component；Ring-shaped guide rail is equipped with the external groove positioned at the inner groovy of inner ring and positioned at outer ring, and movable plate is caught in interior
Between groove and external groove, sliding drive component driving movable plate is slided along ring-shaped guide rail；Laser pen is fixed on chaining part,
Chaining part is connect with movable plate rotary type, and wobble drive Component driver laser pen is swung relative to movable plate.
As a preferred embodiment, movable plate is semi arch movable plate, both ends are equipped with vertical mounting plate；On every piece of mounting plate
It is respectively mounted a chaining part, a laser pen is respectively arranged on each chaining part.
As a preferred embodiment, mounting plate or chaining part are equipped with the graduation mark of display laser pen pivot angle.
As a preferred embodiment, CCD camera, ring laser sensor, LED light are set gradually from top to bottom.
A kind of image-recognizing method based on monocular vision and laser sensing being based on monocular vision and laser using one kind
The pattern recognition device of sense, includes the following steps：S101 acquires the face profile of object by CCD camera, and computer calculates
Whether object meets grasping condition；S102 is swung and sliding laser pen, and laser point is got in the face profile of object, obtains perpendicular
The upward depth information of histogram obtains the coordinate of laser point to carry out coordinate analysis；S103 is asked according to the coordinate of laser point
The normal vector for obtaining the plane that these laser points are constituted, that is, acquire the normal vector of the upper side plane of object, and transfer data to
It is analyzed in computer.
As a preferred embodiment, in step S102, laser pen is swung and slided, laser point is made to fall in body surface difference position
It sets, body surface is identified by the laser point that ring laser sensor acquires, by constantly adjusting laser pen in different positions,
So that laser point caused by laser pen is fallen on body surface different location, at least to get three different points, pass through three
Different points constitutes a plane, the upper side plane of indicator body.
As a preferred embodiment, in step S102, laser point caused by laser pen is fallen on body surface different location, is needed
It goes to identify different positions by ring laser sensor.
As a preferred embodiment, establishing three-dimensional cartesian coordinate system, making the laser point on object to laser pen carries out coordinate score
Analysis, it is straight down z-axis to obtain laser point coordinates by establishing using CCD camera as origin；Pass through CCD camera and object
The height on surface, laser pen extrapolate the coordinate of laser point away from initial point distance.
As a preferred embodiment, in step S103, by the laser point coordinates obtained, one is constituted according to three differences
Then plane solves the normal vector of the plane by mathematical computations；By transmission path by the data transmission of normal vector to meter
In calculation machine, then analyzed.
The principle of the present invention is：
The present invention acts on upper and lower side as plane, and the object that two planes are mutually parallel.Utilize ring laser sensor
Acquisition depth information is got ready in body surface, forms a plane parallel with body surface, and calculate and beat in body surface
World coordinates on point, by being transferred in computer, and then by calculating the normal vector in upper side plane and being transferred to
In computer, computer is made to identify the spatial position of object.And then realize design simple and quickization, working efficiency higher.
By in the three-dimensional system of coordinate established, by the height of known CCD camera to object, the half of ring-shaped guide rail
Diameter, the position for the angle and laser pen that laser pen is swung can be in the hope of the coordinate of a, b, a1, and then calculates the flat of aba1 compositions
The normal vector in face (i.e. the surface of object), is then transferred to computer, and computer is by analysis, where then identifying object
Position.
Generally speaking, the invention has the advantages that：
1. the present invention first is using fairly simple recognition methods, merely by the method for calculating normal vector,
Easy to operate, identification is accurate and recognition speed is fast, and work efficiency is high, significantly reduces the working strength of worker.
2. using image recognition of the present invention, and then normal vector is calculated, the object method that then manipulator is identified and captured,
It is suitable for most the flow work so that company reduces cost of investment, and the cost of manufacture of the present invention is relatively low.
3. the present invention is by practical operation, truly feasible, and recognition speed is fast, and accuracy of identification is relatively high, can be accurate
Crawl object be placed on designated position.
Description of the drawings
Fig. 1 is a kind of flow chart of the image-recognizing method based on monocular vision and laser sensing.
Fig. 2 is a kind of structural schematic diagram of the pattern recognition device based on monocular vision and laser sensing.
Fig. 3 is overall plan implementation figure designed according to this invention.
Fig. 4 is the stereogram of ring laser sensor.
Fig. 5 is the local part drawing of ring laser sensor.
Fig. 6 is the partial enlarged view of ring laser sensor.
Wherein, 1 is CCD camera, and 2 be ring laser sensor, and 3 be movable plate, and 4 and 5 be laser pen, and 6 be chaining part,
7 be holder, and 8 be objective table, and 9 be object, and 10 be computer, and 21 be guide rail outer shroud, and 22 be guide rail inner ring, and 23 be external groove, 24
It is mounting plate for inner groovy, 31,32 be bolt hole, and 61 be bolt hole, and 62 be screw, and 63 be angle indicator, and 64 be fastener.
A, B, A1, B1 are laser pen position, and correspondence irradiates four laser points a, b, a1, b1.L is laser pen to institute
The distance of laser point is beaten, H is distance of the laser point to coordinate origin in z-axis parallel direction.
Specific implementation mode
It is next below that the present invention will be further described in detail.
As shown in Figure 1, this figure, which describes in detail, realizes that monocular vision is quickly identified the process of object and swashed using annular
One introduction of plane normal vector where optical sensor calculates upper side.
S101, CCD camera are connect with ring laser sensor by holder, and CCD camera obtains object plane first
Two-dimensional silhouette, ring-shaped guide rail and semicircle movable plate are nested encryptions, and two swingable laser pens are by indicating graduation mark
Rectangle chaining part connects, and acquires object plane profile by CCD camera and calculates profile size, whether judgment object, which meets, is grabbed
Take condition.Then object is transferred to the ring laser sensor above camera in the two-dimensional coordinate of plane.
S102, after the plane coordinates for obtaining laser pen, laser pen realizes sliding in the plane, coordinates the pendulum of laser pen
It is dynamic, laser point is got in contour of object, and calculate world coordinate system, beating the laser point on object to laser pen carries out coordinate
Analysis.
S103, according to the laser point coordinates obtained, acquiring the plane that these different laser points are constituted, (i.e. object is upper
Side surface) normal vector；It is programmed into row data analysis by existing, to which the normal vector of object is defeated relative to world coordinates
Go out.
Its Computer constantly finds optimal scheme while analysis.
As shown in Fig. 2, by building experiment porch, the placement position of all parts can be clearly seen, needed for experiment
Thing it is less, greatly reduce cost.Simultaneously it can be seen that the implementation of the program can greatly reduce the production and operation of producer
Cost.
As Figure 4-Figure 6, ring laser sensor includes ring-shaped guide rail, movable plate, chaining part, laser pen, sliding driving
Component, wobble drive component；Ring-shaped guide rail is equipped with the external groove positioned at the inner groovy of inner ring and positioned at outer ring, and movable plate is caught in interior
Between groove and external groove, sliding drive component driving movable plate is slided along ring-shaped guide rail；Laser pen is fixed on chaining part,
Chaining part is connect with movable plate rotary type, and wobble drive Component driver laser pen is swung relative to movable plate.Movable plate is semicircle
Arc movable plate, both ends are equipped with vertical mounting plate；It is respectively mounted a chaining part on every piece of mounting plate, is filled on each chaining part
There is a laser pen.
By the analysis of image, the then analysis according to computer to data can find out the normal vector of the body surface,
And then can be for transmission in computer, computer identifies the position where object in turn.Slide drive component, wobble drive component
Preferably driven using motor.
By contrast test, the vision module cost on assembly line is greatly reduced using the present invention, material is main
Including ring-shaped guide rail, movable plate, laser pen；It is general to identify that the speed of three-dimensional workpiece computer is using image processing module method
Between 0.82 second to 0.90 second, because of image processing module, general processing step is to need to carry out gray value processing, threshold value point
It cuts, image noise reduction, calculate workpiece profile etc., and the present invention then only needs simple mathematics geometry calculation formula, it is verified, from swashing
Depth information is transferred to computer and starts to calculate by light pen, between identification workpiece normal vector only needs 0.56 second to 0.66 second.
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, it is other it is any without departing from the spirit and principles of the present invention made by changes, modifications, substitutions, combinations, simplifications,
Equivalent substitute mode is should be, is included within the scope of the present invention.
Claims (10)
1. a kind of pattern recognition device based on monocular vision and laser sensing, it is characterised in that：Including holder, LED light, CCD
Camera, computer, ring laser sensor；Holder is erected at the surface of object under test, the LED illuminated to object
Lamp, the CCD camera, the ring laser sensor that are shot to object are installed on holder；Ring laser sensor includes
The laser pen that can be also slid along annulus can be swung, to which laser pen forms at least three laser points in body surface；It calculates
Machine obtains position and the pivot angle information of laser pen, and the location information of laser point is obtained by CCD camera, finally obtains object
Spatial positional information.
2. a kind of pattern recognition device based on monocular vision and laser sensing described in accordance with the claim 1, it is characterised in that：
Ring laser sensor includes ring-shaped guide rail, movable plate, chaining part, laser pen, sliding drive component, wobble drive component；Ring
Shape guide rail is equipped with the external groove positioned at the inner groovy of inner ring and positioned at outer ring, and movable plate is caught between inner groovy and external groove, sliding
Dynamic drive component driving movable plate is slided along ring-shaped guide rail；Laser pen is fixed on chaining part, and chaining part is rotated with movable plate
Formula connects, and wobble drive Component driver laser pen is swung relative to movable plate.
3. a kind of pattern recognition device based on monocular vision and laser sensing according to claim 2, it is characterised in that：
Movable plate is semi arch movable plate, and both ends are equipped with vertical mounting plate；A chaining part is respectively mounted on every piece of mounting plate, each
A laser pen is respectively arranged on chaining part.
4. a kind of pattern recognition device based on monocular vision and laser sensing described in accordance with the claim 3, it is characterised in that：
Mounting plate or chaining part are equipped with the graduation mark of display laser pen pivot angle.
5. a kind of pattern recognition device based on monocular vision and laser sensing according to claim 2, it is characterised in that：
CCD camera, ring laser sensor, LED light are set gradually from top to bottom.
6. a kind of image-recognizing method based on monocular vision and laser sensing, described in any one of claim 1-5
A kind of pattern recognition device based on monocular vision and laser sensing, it is characterised in that：Include the following steps：
S101 acquires the face profile of object by CCD camera, and computer calculates whether object meets grasping condition；
S102 is swung and sliding laser pen, and laser point is got in the face profile of object, obtains the depth letter on vertical direction
Breath, to carry out coordinate analysis, obtains the coordinate of laser point；
S103 acquires the normal vector for the plane that these laser points are constituted, that is, acquires the upside of object according to the coordinate of laser point
The normal vector of plane, and transfer data in computer and analyzed.
7. a kind of image-recognizing method based on monocular vision and laser sensing according to claim 6, it is characterised in that：
In step S102, laser pen is swung and slided, so that laser point is fallen in body surface different location, is adopted by ring laser sensor
The laser point identification body surface of collection makes laser point caused by laser pen by constantly adjusting laser pen in different positions
It falls on body surface different location, at least to get three different points, constitute a plane by three different points, refer to
Show the upper side plane of object.
8. a kind of image-recognizing method based on monocular vision and laser sensing according to claim 7, it is characterised in that：
In step S102, laser point caused by laser pen is fallen on body surface different location, need to be gone by ring laser sensor
Identify different positions.
9. a kind of image-recognizing method based on monocular vision and laser sensing according to claim 6, it is characterised in that：
Three-dimensional cartesian coordinate system is established, beating the laser point on object to laser pen carries out coordinate analysis, obtains laser point coordinates, passes through
It establishes using CCD camera as origin, is straight down z-axis；By the height of CCD camera and body surface, laser pen is away from original
Point distance, extrapolates the coordinate of laser point.
10. a kind of image-recognizing method based on monocular vision and laser sensing, feature exist according to claim 6
In：In step S103, by the laser point coordinates obtained, a plane is constituted according to three differences, then passes through mathematics
Calculate the normal vector for solving the plane；By transmission path by the data transmission to computer of normal vector, then divided
Analysis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810171987.1A CN108548482B (en) | 2018-03-01 | 2018-03-01 | A kind of pattern recognition device and method based on monocular vision and laser sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810171987.1A CN108548482B (en) | 2018-03-01 | 2018-03-01 | A kind of pattern recognition device and method based on monocular vision and laser sensing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108548482A true CN108548482A (en) | 2018-09-18 |
| CN108548482B CN108548482B (en) | 2019-11-15 |
Family
ID=63516013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810171987.1A Active CN108548482B (en) | 2018-03-01 | 2018-03-01 | A kind of pattern recognition device and method based on monocular vision and laser sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108548482B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109541626A (en) * | 2018-12-12 | 2019-03-29 | 华南农业大学 | Objective plane normal direction amount detecting device and detection method |
| CN110321836A (en) * | 2019-07-01 | 2019-10-11 | 芜湖启迪睿视信息技术有限公司 | A kind of conveying material detection method based on image and laser point cloud atlas |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130323434A1 (en) * | 2012-05-29 | 2013-12-05 | King Abdullah University Of Science And Technology | Inkjet printing with in situ fast annealing for patterned multilayer deposition |
| CN103791838A (en) * | 2014-03-07 | 2014-05-14 | 中国工程物理研究院激光聚变研究中心 | Method for automatic measurement of pose of micro part |
| CN104034261A (en) * | 2014-06-10 | 2014-09-10 | 深圳先进技术研究院 | Surface normal measurement device and method |
| CN105021124A (en) * | 2015-04-16 | 2015-11-04 | 华南农业大学 | Planar component three-dimensional position and normal vector calculation method based on depth map |
| CN107330936A (en) * | 2017-06-21 | 2017-11-07 | 清华大学深圳研究生院 | A kind of Double Circle mark localization method and system based on monocular vision |
-
2018
- 2018-03-01 CN CN201810171987.1A patent/CN108548482B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130323434A1 (en) * | 2012-05-29 | 2013-12-05 | King Abdullah University Of Science And Technology | Inkjet printing with in situ fast annealing for patterned multilayer deposition |
| CN103791838A (en) * | 2014-03-07 | 2014-05-14 | 中国工程物理研究院激光聚变研究中心 | Method for automatic measurement of pose of micro part |
| CN104034261A (en) * | 2014-06-10 | 2014-09-10 | 深圳先进技术研究院 | Surface normal measurement device and method |
| CN105021124A (en) * | 2015-04-16 | 2015-11-04 | 华南农业大学 | Planar component three-dimensional position and normal vector calculation method based on depth map |
| CN107330936A (en) * | 2017-06-21 | 2017-11-07 | 清华大学深圳研究生院 | A kind of Double Circle mark localization method and system based on monocular vision |
Non-Patent Citations (3)
| Title |
|---|
| 唐宇等: "平面工件的识别与定位方法研究", 《机械设计与制造》 * |
| 王修岩，程婷婷: "基于单目视觉的工业机器人智能抓取研究", 《机械设计与制造》 * |
| 高学海等: "基于单目视觉和激光测距仪的位姿测量算法", 《仪器仪表学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109541626A (en) * | 2018-12-12 | 2019-03-29 | 华南农业大学 | Objective plane normal direction amount detecting device and detection method |
| CN110321836A (en) * | 2019-07-01 | 2019-10-11 | 芜湖启迪睿视信息技术有限公司 | A kind of conveying material detection method based on image and laser point cloud atlas |
| CN110321836B (en) * | 2019-07-01 | 2023-07-21 | 芜湖启迪睿视信息技术有限公司 | Conveyed material detection method based on image and laser point cloud image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108548482B (en) | 2019-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110509300B (en) | Steel hoop processing and feeding control system and control method based on three-dimensional visual guidance | |
| CN110434671B (en) | Cast member surface machining track calibration method based on characteristic measurement | |
| CN108182689B (en) | Three-dimensional identification and positioning method for plate-shaped workpiece applied to robot carrying and polishing field | |
| CN107186548B (en) | A kind of five-axle number control machine tool rotating shaft geometric error detection method | |
| CN201653373U (en) | Triaxial non-contact image measuring system | |
| CN109612390B (en) | Large-size workpiece automatic measuring system based on machine vision | |
| KR102056664B1 (en) | Method for work using the sensor and system for performing thereof | |
| CN103433810B (en) | Complicated curve surface normal vector on-machine detection device and method | |
| CN101839700A (en) | Non-contact image measuring system | |
| CN113465511B (en) | Steel coil size online measurement and omnibearing end surface defect online detection method | |
| CN102506830B (en) | Vision-based positioning method and device | |
| KR102634535B1 (en) | Method for recognizing touch teaching point of workpiece using point cloud analysis | |
| CN107588723B (en) | Circular mark leak source detection method on a kind of High-speed target based on two-step method | |
| CN104217441A (en) | Mechanical arm positioning fetching method based on machine vision | |
| CN106643545A (en) | Calibration method for steel rail profile measured by adopting laser displacement technology | |
| CN108548482B (en) | A kind of pattern recognition device and method based on monocular vision and laser sensing | |
| CN106767401A (en) | A kind of shaft hole series part based on cross laser and machine vision determines appearance localization method | |
| CN109341532A (en) | A kind of part coordinate scaling method based on structure feature towards automatic assembling | |
| CN111531407B (en) | Workpiece attitude rapid measurement method based on image processing | |
| US11816754B2 (en) | Measurement parameter optimization method and device, and computer control program stored on computer-readable storage medium | |
| CN103335602A (en) | Multistation and multiparameter high precision measurement method and device of light curtain shaft part | |
| CN110081821A (en) | Intelligent high-speed rail white body assembling quality detection device and its method | |
| CN115619738A (en) | Detection method for module side seam welding after welding | |
| CN109405767B (en) | Casting head cutting track determination method based on laser profile measuring instrument | |
| CN110568866B (en) | Three-dimensional curved surface visual guidance alignment system and alignment 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 |