CN107016699A - A kind of color coding approach of the variegated particle of automatic identification - Google Patents
A kind of color coding approach of the variegated particle of automatic identification Download PDFInfo
- Publication number
- CN107016699A CN107016699A CN201710148662.7A CN201710148662A CN107016699A CN 107016699 A CN107016699 A CN 107016699A CN 201710148662 A CN201710148662 A CN 201710148662A CN 107016699 A CN107016699 A CN 107016699A
- Authority
- CN
- China
- Prior art keywords
- particle
- color
- hsi
- object chain
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- 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
- G06T2207/10024—Color image
Abstract
The present invention relates to a kind of color coding approach of the variegated particle of automatic identification.This method has tri- values of R, G, B using colored area array cameras record particle stream picture, each pixel.It is HSI color model first by the RGB three primary colors model conversion of acquisition, HSI models are a cylinders based on cylinder polar coordinate system, HS can be reduced to two dimensional surface, I is third dimension information.So just with a rectangle frame a kind of color can be defined in HS planes, add the Intensity of the third dimension, impurity block region can be defined as a cuboid.The pixel in the range of the picture that collects is identified with the cuboid of definition, the profile, centroid position coordinate and size of impurity block is judged.This method can quickly and accurately detect the information such as color, number, the size of coloring impurities in particle flux, liberate labour, with great engineering application value.
Description
Technical field
The present invention relates to a kind of detection method for being used to detect variegated particle in particle flux, particularly a kind of automatic identification is miscellaneous
The color coding approach of colored particle, belongs to field of photodetection.
Background technology
High polymer product is a kind of rapidoprint that China develops rapidly in recent years, is widely used in each neck
Domain, turns into important material in mechanical fitting, toy for children, industrial or agricultural, biomedicine and packaging industry.When with industry
During large-scale production, material zero defect is not only an aesthetic problem, as a rule or an absolute quality requirements.
Therefore, its raw material supplier constantly increases to the quality requirement of product.
The country is used for colour particles scanning and automatically record specified defect color, appearance frequency also without complete set at present
The system of the relevant statistics such as rate, import equipment is relied primarily on to the test in laboratory of polymer industry raw material.
The image of particle flux is recorded with colored CCD area array cameras, image is analyzed, particle is quickly and accurately obtained
The information such as the color of coloring impurities, number, size, has liberated labour in stream, with great engineering application value.
The content of the invention
It is an object of the invention to the deficiency existed for prior art, there is provided a kind of colour of the variegated particle of automatic identification
Coloring impurities color, corresponding number and size in coding method, automatic detection particle flux, this method are accurately easy, practical.
To reach above-mentioned purpose, the present invention uses following technical proposals：
A kind of color coding approach of the variegated particle of automatic identification, it is characterised in that operating procedure is as follows：
1) using colored area array cameras record particle stream picture, each pixel has tri- values of R, G, B.
2) by step 1) in obtain image HSI is converted to by RGB.The each pixel of original image of cameras record has
Tri- values of R, G, B, RGB models are (as shown in Figure 2) to be based on the principle of three primary colours, towards hardware, is easy to the collection and display of color, but
It is not easy to image procossing.So needing RGB models to HSI (colourity Hue, saturation degree Saturation, brightness Intensity)
Model (as shown in Figure 3) is changed.This is actually that a cube based on cartesian coordinate system is based on cylinder to one
The conversion of the cylinder of polar coordinate system.
The polar definition of HSI cylinders is such, radial distribution of the brightness along cylinder, and center is on the outside of black (0%)
For white (100%).Saturation degree is distributed along the height of cylinder, from bottom to top from 0% (most weak) to 100% (most strong).Colourity edge
Central shaft is distributed into 0-360 °.To 60 ° (Huang), 120 ° (green), 180 ° (green grass or young crops), 240 ° (blueness), 300 ° of (product since 0 ° (red)
It is red).By HSI models outer surface, i.e. I=1 unfolded surface, as shown in Figure 4.
Rgb value is normalized into this H-S plane (I=1)：
Conversion segmented model after normalization is expressed as follows (RGB → HSI)：
S=1-Min
I=Max
HSI models are the description of defective particles impurity block relative to the advantage of traditional RGB models, as shown in Figure 5.Often
The point of individual brown represents a pixel on brown impurity block, and the point of each white represents a pixel of background, for simplification
Here bidimensional is only shown.
In R/G planes, it is difficult to which define one does not only include background (white point) comprising impurity block (point of brown)
Region.If plus 3-dimensional B, describing such a color region will become hardly possible.And in HS planes, brown point
Region can be very easily with a rectangle frame：Hue=10~50 °, Saturation=20~100% goes definition.Add
3-dimensional Intensity, the region of impurity block is changed into cuboid from rectangle frame, is not still a complicated shape.It is such as dark brown
Color can use Hue=10~50 °, and Saturation=20~100%, Intensity=0~50% goes to represent.HSI models
Advantage is apparent.
3) Hue, Saturation are utilized, Intensity scope defines the impurity classification to be recognized.
4) for HSI in step 3) defined in the range of pixel recognized, be judged as impurity block, and calculate this
Profile, centroid position coordinate and the size of impurity block.Wherein centroid position coordinate (x, y) and size (size) store following (n
Represent colour type)：
During Particles Moving, camera records particle stream picture in real time, and same particle can be detected repeatedly, use
Following tracing algorithm avoids the repeat count to same particle.
5) color, centroid position coordinate are stored in object chain, are designated as：
6) new present frame per treatment, repeat step 2), step 4), obtain defective particles impurity block in current frame image
Color, position coordinates and size it is as follows：
7) by step 6) in position coordinates successively with step 5) in each element of object chain be compared, if there is one
Individual color is identical, and position falls in object chain in the range of the course of element (i), and apart from more than a certain threshold value (Min_
distance).I.e. centroid position coordinate meets following relation：
Wherein：X_d=detection [n] .centre [l] .x-MB [n] .centre [i] .x
Y_d=detection [n] .centre [l] .y-MB [n] .centre [i] .y
It is exactly the moving region of the element (i) of former frame then to think the position, and the element for substituting object chain is updated with it
(i) number of times, tracked increases by 1 time.
When the number of times that a certain element is tracked in object chain accumulates to certain number of times, then it is assumed that the defect tracking into
Work(, defect number increase by 1.
When step 6) in all position coordinates inspections finish, there is the position not with object chain Match of elemental composition, then added
Enter to object chain, as new object element.
If there is the element for failing matching in object chain, the number of times that the element is not tracked increases by 1 time.
When the number of times that a certain element is not tracked in object chain accumulates to certain number of times, then it is assumed that the target from
Field range is opened, the target chain element is emptied, no longer follows the trail of next time.
Implementation procedure is as shown in Figure 6.Particle right-to-left is moved, 3 particles (red, green, blue) the in the first two field picture
Once be detected, respectively each particle periphery define a rectangle frame, when it defines next image frame grabber this
The transportable scope of particle.When next two field picture is collected, particle moves a certain distance, all grains being detected
Rectangle frame of the son respectively with previous definition is compared.Three particles are successfully compared, and the distance of red and green particle movement is big
In threshold value (Min_distance), follow the trail of successfully.Blue particles are not almost moved, and are judged as background stain.
8) new present frame is received, above step is repeated.
Brief description of the drawings
Fig. 1 is the flowsheet of the present invention.
Fig. 2 is traditional RGB color spatial model.
Fig. 3 is the HSI column color space schematic diagrames that the present invention is used.
Fig. 4 is the result that rgb value is normalized to this H-S plane.
Fig. 5 is superiority of the HSI column color spaces of the invention utilized compared with rgb color space.
Fig. 6 is tracing algorithm process schematic of the present invention.
Fig. 7 is the present frame that colored area array cameras records real-time particle stream picture.
Fig. 8 is the HIS color space distribution schematic diagram of two kinds of defective particles classifications defined in the present invention.
Fig. 9 is the result of current frame image.
Embodiment
Details are as follows for the preferred embodiments of the present invention combination accompanying drawing：
A kind of color coding approach of the variegated particle of automatic identification, fast and accurately counts black and coloured silk in particle flux respectively
Colored foreign number, operating procedure is as follows：
1) using colored area array cameras record particle stream picture, present frame is as shown in Figure 7.
2) by step 1) in obtain image HSI color spaces are converted to by RGB.Specific algorithm is as follows：
Rgb value is normalized into this H-S plane (I=1)：
Then there is following 6 kinds of situations：
1. R=1, G >=B；
2. R=1, G ＜ B；
3. G=1, R ＞ B；
4. G=1, R≤B；
5. B=1, R ＜ G；
6. B=1, R >=G.
Conversion segmented model after normalization is expressed as follows (RGB → HSI)：
S=1-Min
I=Max
3) the two kinds of impurity to be recognized classifications are defined：Black and other.As shown in Figure 8.
Black particles are defined as：H=0~360 °, S=0~40%, I=0~55%.In the definition of this colour type,
As long as the particle that brightness value is not higher than 55% will be identified, and not mind that it is dark blue or dark red or dark-grey.
The impurity of other classifications is defined as：H=0~360 °, S=30~100%, I=0~100%.In this colour type
Definition in, as long as the particle that intensity value is not less than 40% will be identified as the impurity of other classifications, it is dark red not distinguish
Or it is pale red or pale yellow.
4) according to step 3) in impurity definition, process step 2) in obtain image, obtain impurity block on defective particles
Profile, centroid position and size.As a result it is as shown in Figure 9.
5) color, centroid position information are stored in object chain.
6) new present frame per treatment, repeat step 2), step 4), obtain color, centroid position, the defect of present frame
Size.
7) by step 6) in position coordinates successively with 5) in each element of object chain be compared, if there is a face
Color is identical, and position falls in object chain in the range of the course of element (i), and apart from more than a certain threshold value (Min_
distance).I.e. centroid position coordinate meets following relation：
Wherein：X_d=detection [n] .centre [l] .x-MB [n] .centre [i] .x
Y_d=detection [n] .centre [l] .y-MB [n] .centre [i] .y
It is exactly the moving region of the element (i) of former frame then to think the position, and the element for substituting object chain is updated with it
(i) number of times, tracked increases by 1 time.
When the number of times that a certain element is tracked in object chain accumulates to certain number of times, then it is assumed that the defect tracking into
Work(, defect number increase by 1.
When all position coordinates inspections are finished in present frame, the position not with object chain Match of elemental composition is still suffered from, then by it
Object chain is added to, as new object element.
If there is the element for failing matching in object chain, the number of times that the element is not tracked increases by 1 time.
When the number of times that a certain element is not tracked in object chain accumulates to certain number of times, then it is assumed that the target from
Field range is opened, the target chain element is emptied, no longer follows the trail of next time.
8) new present frame is received, above step is repeated.
Claims (1)
1. a kind of color coding approach of the variegated particle of automatic identification, it is characterised in that operating procedure is as follows：
1) using colored area array cameras record particle stream picture, each pixel has tri- values of R, G, B；
2) by step 1) in obtain image HSI is converted to by RGB；The each pixel of original image of cameras record have R, G,
The value of B tri-, RGB models are based on the principle of three primary colours, towards hardware, are easy to the collection and display of color, but are not easy to image procossing,
So to HSI --- colourity Hue, saturation degree Saturation, brightness Intensity ---, model is changed by RGB models；
This is actually conversion of the cube based on cartesian coordinate system to a cylinder based on cylinder polar coordinate system；
HSI cylinders polar coordinates are defined：Radial distribution of the brightness along cylinder, center is for white (100%) on the outside of black (0%)；It is full
Short transverse with degree along cylinder is distributed, from bottom to top from most weak (0%) to most strong (100%)；Colourity along central shaft into
0-360 ° of distribution；To yellow (60 °), green (120 °), blue or green (180 °), blue (240 °), pinkish red (300 °) since red (0 °)；By HSI
The unfolded surface of model outer surface, i.e. I=1, obtains H-S planes；
Rgb value is normalized into H-S planes (I=1)：
Conversion segmented model after normalization is expressed as follows：RGB→HSI
S=1-min
I=max
HSI models are the description of defective particles impurity block relative to the advantage of traditional RGB models, and the point of each brown is represented
A pixel on brown impurity block, the point of each white represents a pixel of background, in order to which simplification only shows bidimensional here；
In R/G planes, it is difficult to defining one only includes impurity block --- the point of brown, and not comprising background --- white point
Region；And in HS planes, the region of brown point is with a rectangle frame：Hue=10~50 °, Saturation=20~100% is gone
Definition；Plus 3-dimensional Intensity, the region of impurity block is changed into cuboid from rectangle frame, dark-brown Hue=10~50 °,
Saturation=20~100%, Intensity=0~50% goes to represent；
3) Hue, Saturation are utilized, Intensity scope defines the impurity classification to be recognized；
4) for HSI in step 3) defined in the range of pixel recognized, be judged as impurity block, and calculate the impurity
Profile, centroid position coordinate and the size of block；Wherein centroid position coordinate x, y and size size storage are as follows, and n represents color
Classification：
During Particles Moving, camera records particle stream picture in real time, and same particle can be detected repeatedly, using following
Tracing algorithm avoids the repeat count to same particle；
5) color, centroid position coordinate are stored in object chain, are designated as：
6) new present frame per treatment, repeat step 2), step 4), obtain the face of defective particles impurity block in current frame image
Color, position coordinates and size are as follows：
7) by step 6) in position coordinates successively with step 5) in each element of object chain be compared, if there is a face
Color is identical, and position falls in object chain in the range of element i course, and apart from more than a certain threshold value Min_
distance；I.e. centroid position coordinate meets following relation：
Wherein：X_d=detection [n] .centre [l] .x-MB [n] .centre [i] .x
Y_d=detection [n] .centre [l] .y-MB [n] .centre [i] .y
It is exactly the element i of former frame moving region then to think the position, and substitutes the element i of object chain with its renewal, is followed the trail of
The number of times arrived increases by 1 time；
When the number of times that a certain element is tracked in object chain accumulates to certain number of times, then it is assumed that defect tracking success, lack
Fall into number increase by 1；
When step 6) in all position coordinates inspections finish, there is the position not with object chain Match of elemental composition, then add it to
Object chain, as new object element；
If there is the element for failing matching in object chain, the number of times that the element is not tracked increases by 1 time；
When the number of times that a certain element is not tracked in object chain accumulates to certain number of times, then it is assumed that the target has been moved off regarding
Wild scope, empties the target chain element, no longer follows the trail of next time；
In tracing algorithm procedure chart, the movement of particle right-to-left, 3 particle-red, green, blue-the first in the first two field picture
It is secondary to be detected, define a rectangle frame, this grain when it defines next image frame grabber in each particle periphery respectively
The transportable scope of son；When next two field picture is collected, particle moves a certain distance, all particles being detected
The rectangle frame with previous definition is compared respectively；Three particles are successfully compared, and the distance of red and green particle movement is more than
Threshold value (Min_distance), is followed the trail of successfully；Blue particles are not almost moved, and are judged as background stain；
8) new present frame is received, above step is repeated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710148662.7A CN107016699A (en) | 2017-03-14 | 2017-03-14 | A kind of color coding approach of the variegated particle of automatic identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710148662.7A CN107016699A (en) | 2017-03-14 | 2017-03-14 | A kind of color coding approach of the variegated particle of automatic identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107016699A true CN107016699A (en) | 2017-08-04 |
Family
ID=59440847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710148662.7A Pending CN107016699A (en) | 2017-03-14 | 2017-03-14 | A kind of color coding approach of the variegated particle of automatic identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107016699A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108765504A (en) * | 2018-05-29 | 2018-11-06 | 上海视疆科学仪器有限公司 | A kind of method of variegated particle in quick screening high-velocity particles stream |
| CN109270952A (en) * | 2018-09-19 | 2019-01-25 | 清远市飞凡创丰科技有限公司 | A kind of agricultural land information acquisition system and method |
| CN110296691A (en) * | 2019-06-28 | 2019-10-01 | 上海大学 | Merge the binocular stereo vision measurement method and system of IMU calibration |
| CN110599555A (en) * | 2019-09-24 | 2019-12-20 | 西南大学 | Plasma nanoparticle dark field microscopic imaging analysis method based on HSI color coding |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398392A (en) * | 2007-09-26 | 2009-04-01 | 中国科学院自动化研究所 | Cotton impurity high speed real-time detection method based on HSI color space |
| CN105930833A (en) * | 2016-05-19 | 2016-09-07 | 重庆邮电大学 | Vehicle tracking and segmenting method based on video monitoring |
| CN106327528A (en) * | 2016-08-23 | 2017-01-11 | 常州轻工职业技术学院 | Moving object tracking method and operation method of unmanned aerial vehicle |
-
2017
- 2017-03-14 CN CN201710148662.7A patent/CN107016699A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398392A (en) * | 2007-09-26 | 2009-04-01 | 中国科学院自动化研究所 | Cotton impurity high speed real-time detection method based on HSI color space |
| CN105930833A (en) * | 2016-05-19 | 2016-09-07 | 重庆邮电大学 | Vehicle tracking and segmenting method based on video monitoring |
| CN106327528A (en) * | 2016-08-23 | 2017-01-11 | 常州轻工职业技术学院 | Moving object tracking method and operation method of unmanned aerial vehicle |
Non-Patent Citations (5)
| Title |
|---|
| LI, YJ 等: "Key Technology in Detecting and Eliminating Isomerism Fibre in Cotton", 《ICEMI 2007: PROCEEDINGS OF 2007 8TH INTERNATIONAL CONFERENCE ON ELECTRONIC MEASUREMENT & INSTRUMENTS》 * |
| MARCHAL, PC 等: "Expert system based on computer vision to estimate the content of impurities in olive oil samples", 《JOURNAL OF FOOD ENGINEERING》 * |
| 栾悉道 等: "《多媒体情报处理技术》", 31 May 2016, 国防工业出版社 * |
| 赵谦 等: "《智能视频图像处理技术与应用》", 30 September 2016, 西安电子科技大学出版社 * |
| 高伟 等: "基于HSI颜色空间的棉花杂质高速实时检测方法", 《自动化学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108765504A (en) * | 2018-05-29 | 2018-11-06 | 上海视疆科学仪器有限公司 | A kind of method of variegated particle in quick screening high-velocity particles stream |
| CN109270952A (en) * | 2018-09-19 | 2019-01-25 | 清远市飞凡创丰科技有限公司 | A kind of agricultural land information acquisition system and method |
| CN110296691A (en) * | 2019-06-28 | 2019-10-01 | 上海大学 | Merge the binocular stereo vision measurement method and system of IMU calibration |
| CN110599555A (en) * | 2019-09-24 | 2019-12-20 | 西南大学 | Plasma nanoparticle dark field microscopic imaging analysis method based on HSI color coding |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107016699A (en) | A kind of color coding approach of the variegated particle of automatic identification | |
| CN110415181B (en) | Intelligent identification and grade judgment method for RGB (red, green and blue) images of flue-cured tobacco in open environment | |
| Liu et al. | A method of segmenting apples at night based on color and position information | |
| Lee et al. | Rapid color grading for fruit quality evaluation using direct color mapping | |
| CN106384117B (en) | A kind of vehicle color identification method and device | |
| Chen et al. | A YOLOv3-based computer vision system for identification of tea buds and the picking point | |
| CN108596880A (en) | Weld defect feature extraction based on image procossing and welding quality analysis method | |
| CN108537239A (en) | A kind of method of saliency target detection | |
| CN105574514B (en) | The raw tomato automatic identifying method in greenhouse | |
| CN102982350A (en) | Station caption detection method based on color and gradient histograms | |
| CN107301634A (en) | A kind of robot automatic sorting method and system | |
| CN110687121B (en) | Intelligent online detection and automatic grading method and system for ceramic tiles | |
| EP2807634B1 (en) | Method for preparing a system which is used to optically identifying objects, laboratory image capturing system for carrying out such a method, and arrangement comprising the laboratory image capturing system and the system | |
| CN109684971A (en) | Magic cube-solving robot algorithm executes method | |
| CN108480223A (en) | A kind of workpiece sorting system and its control method | |
| CN106903075A (en) | A kind of vamp logo multi-directional visions detection method and system | |
| CN104952754A (en) | Coated silicon chip sorting method based on machine vision | |
| Bulanon et al. | A machine vision system for the apple harvesting robot | |
| CN108830908A (en) | A kind of magic square color identification method based on artificial neural network | |
| CN106558044A (en) | The resolution measuring method of image module | |
| CN109087287A (en) | Method based on image analysis evaluation space cabin material surface bacterial plaque elimination effect | |
| CN210377552U (en) | Fruit is multiaspect image acquisition device for classification | |
| Yao et al. | Study on detection method of external defects of potato image in visible light environment | |
| CN115601690A (en) | Edible fungus environment detection method based on intelligent agriculture | |
| CN109829496A (en) | A kind of physical measurement classification method and equipment |
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 |