CN106124511A - A kind of fruit surface defect detection method returning rectification based on adaption brightness - Google Patents
A kind of fruit surface defect detection method returning rectification based on adaption brightness Download PDFInfo
- Publication number
- CN106124511A CN106124511A CN201610349350.8A CN201610349350A CN106124511A CN 106124511 A CN106124511 A CN 106124511A CN 201610349350 A CN201610349350 A CN 201610349350A CN 106124511 A CN106124511 A CN 106124511A
- Authority
- CN
- China
- Prior art keywords
- row
- image
- fruit
- img
- gray
- 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
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a kind of fruit surface defect detection method returning based on adaption brightness and correcting.Fruit RGB color image is converted to initial gray image, segmentation background and target, carry out binary conversion treatment, carry out individual element point with binaryzation fruit image with initial gray image to be multiplied, obtain fruit gray level image, ranks pixel in fruit gray level image is obtained both horizontally and vertically regression figure picture by least squares regression method respectively, fruit gray level image be respectively divided by both horizontally and vertically regression figure picture obtain both horizontally and vertically brightness correction image and carry out arithmetic mean calculate obtain brightness correction image, brightness correction image carries out binaryzation and completes to fill out hole and medium filtering acquisition fruit surface defect image.The present invention detects surface difference brightness defect in the case of overcoming globoid surface brightness uneven;In terms of fruit and quality of agricultural product computer vision on-line checking, there is application potential.
Description
Technical field
The present invention relates to computer visual image processing method, be specifically related to a kind of based on adaption brightness recurrence rectification
Fruit surface defect detection method.
Background technology
Surface defects detection is one of important evidence of fruit grading, has strict in the fruit rating scale of countries in the world
Regulation.The most a large amount of scholar's research detect fruit and surface of agricultural products defect by computer vision mode, but many
Agricultural product are globoids, and the gray value in the middle part of X-Y scheme is significantly larger than the gray value at edge, cause surface defect image to detect
Difficulty.
Finding through the retrieval of existing technology, method is broadly divided into three classes:
1) processing method based on spheroid gray level model.Such as patent documentation Chinese patent CN101984346A describes one
Plant fruit surface defect detection method based on low-pass filtering, first obtain the R component image removing background, utilize fruit colored
Image carries out low-pass filtering then inverse discrete Fourier transform by discrete Fourier transform and changes acquisition surface brightness image, Qian Zhetu
As obtaining uniforming luminance picture divided by latter image, then using single threshold to realize fruit surface defect segmentation, this technology relates to
More complicated Digital Signal Processing increases online development difficulty, and the optimal parameter of low-pass filtering needs along with different luminous environments
Artificial experience is wanted again to debug selection;Chinese patent CN102788806A utilizes fruit RGB image and NIR image, calculates contrast
The defect shape of fruit, size, but fruit is not strict spheroid, and this patent bianry image boundary rectangle Breadth Maximum approximates
Fruit diameter, the half of Breadth Maximum is as iterations end condition.Non-circular oval fruit can be produced by this method
Raw error, and illumination lambert's phenomenon that oval fruit major axis and short axle are subject to is different direct direct with this area pixel point
Carry out brightness average treatment, defects detection can be brought error;Li Jiangbo et al. (2013) utilize illumination mode with
Image is than technology for detection Citrus sinensis Osbeck surface defect, and the method still relates to more complicated Digital Signal Processing to be increased in line development difficulty
Degree, and the optimal parameter of low-pass filtering needs artificial experience again to debug selection along with different luminous environments.(Automatic
detection of common surface defects on oranges using combined lighting
transform and image ratio methods Jiangbo Li,Xiuqin Rao,Postharvest Biology
and Technology 2013)。
2) processing method based on surface texture feature.L ó pez-Garc í a F et al. (2010) utilize multiplex images theoretical
And surface texture feature Algorithm for Training method detection Citrus sinensis Osbeck surface defect, this method comparison complexity is not readily used for online, and
Detection Citrus sinensis Osbeck surface defect limited types.(López-García F,Andreu-García G,Blasco J,et
al.Automatic detection of skin defects in citrus fruits using a multivariate
image analysis approach[J].Computers and Electronics in Agriculture,2010,71
(2):189-19)。
3) processing method based on multi-optical spectrum imaging technology.J.Blascoa et al. utilizes multispectral imaging equipment to carry out umbilicus
Orange surface deficiency analysis, the method hardware cost higher and complicated (2007) (J.Blascoa, N.Aleixos. (2007)
.Citrus sorting by identification of the most common defects using
multispectral computer vision.Journal of Food Engineering 83(2007)384–393)。
There is Processing Algorithm and for on-line checking or rely on relatively costly complex hardware compared with complicated difficult in existing method
The problem of imaging technique, it is therefore desirable to new fruit surface defect detection method.
Summary of the invention
It is an object of the invention to provide a kind of fruit surface defect detection method returning based on adaption brightness and correcting,
Detection method is simpler and applicable on-line checking occasion, and surface defects detection type is more and object practicality is wider.
The step that the present invention solves the used technical scheme of its technical problem is as follows:
1) obtain fruit RGB color image, fruit RGB color image is converted to initial gray image;
2) by initial gray image segmentation background and target, carry out binary conversion treatment and obtain binaryzation fruit image, use 1 table
Showing fruit target, 0 represents background;
3) by step 2) the binaryzation fruit image that obtains and step 1) the initial gray image that obtains carries out individual element
Point is multiplied, and obtains fruit gray level image;
4) all pixels that every a line in fruit gray level image and every string remove background pixel carry out a young waiter in a wineshop or an inn respectively
The polynomial regression Calculation of multiplication quadratic term, obtains horizontal direction regression figure picture and vertical direction regression figure picture respectively;
5) fruit gray level image is respectively divided by horizontal direction regression figure picture and vertical direction regression figure picture acquisition level side
To brightness correction image and vertical direction brightness correction image, by horizontal direction brightness correction image and vertical direction brightness correction
Image addition is averaged acquisition brightness correction image;
6) by segmentation threshold, brightness correction image is carried out binaryzation, then complete to fill out hole and medium filtering obtains fruit
Surface defect image.
Described step 4) particularly as follows:
4.1) start to scan the gray value pixel quantity more than 0 successively more than 3 at fruit gray level image top line
Each row, for every a line, by the gray value of this row gray value pixel more than 0 by from left to right order writing line successively to
Amount f (n);
4.2) as follows row vector f (n) is carried out least square regression calculating, obtain row quadratic term coefficient k2, OK
Monomial coefficient k1With row constant term k0;
f1(n)=k2n2+k1n+k0+ε0 (1)
Wherein, f1(n) grey scale pixel value;The row number at n grey scale pixel value place;k2Row quadratic term coefficient;k1OK
Monomial coefficient;k0Row constant term;ε0Row residual error;
4.3) row residual epsilon is taken again0It is 0, by row quadratic term coefficient k2, row Monomial coefficient k1With row constant term k0, pixel ash
The row n at angle value place substitutes into below equation and calculates acquisition row recurrence gray value f again2(n):
f2(n)=k2n2+k1n+k0+ε0 (2)
Wherein, f2N () returns gray value for row;
4.4) using calculating acquisition row recurrence gray value as the pixel value of this location of pixels in horizontal direction regression figure picture, right
Original gray value is retained as pixel value in the gray value pixel quantity more than the 0 row horizontal direction regression figure picture less than 3;
4.5) start to scan the gray value pixel quantity more than 0 successively more than 3 at fruit gray level image Far Left string
Each row, for every string, the gray value of this row gray value pixel more than 0 is write successively by order from top to bottom row to
Amount g (m);
4.6) as follows column vector g (m) is carried out least square regression calculating, obtain row quadratic term coefficient p2, row
Monomial coefficient p1With row constant term p0;
g1(m)=p2m2+p1m+p0+ε1 (3)
Wherein, g1(m) grey scale pixel value;The line number at m grey scale pixel value place;p2Row quadratic term coefficient;p1Row
Monomial coefficient;p0Row constant term;ε1Row residual error;
4.7) row residual epsilon is taken again1It is 0, by row quadratic term coefficient p2, row Monomial coefficient p1With row constant term p0, pixel ash
The row m at angle value place substitutes into below equation and calculates acquisition row recurrence gray value g again2(m):
g2(m)=p2m2+p1m+p0+ε1 (4)
Wherein, g2M () returns gray value for row;
4.8) using calculating acquisition row recurrence gray value as the pixel value of this location of pixels in vertical direction regression figure picture, right
Original gray value is retained as pixel value in the gray value pixel quantity more than the 0 row vertical direction regression figure picture less than 3.
Described step 5) particularly as follows:
5.1) below equation is used to obtain horizontal direction brightness correction image,
IMGCH=IMGFG/IMGRH (5)
In formula: IMGCHHorizontal direction brightness correction image;IMGRHHorizontal direction regression figure picture;IMGFHFruit ash
Degree image;
5.2) below equation is used to obtain vertical direction brightness correction image,
IMGCV=IMGFG/IMGRV (6)
In formula: IMGCVVertical direction brightness correction image;IMGRVVertical direction regression figure picture;IMGFGFruit ash
Degree image;
5.3) use below equation acquisition brightness correction image:
IMGCorrect=0.5 × MGCH+0.5×IMGCV (7)
In formula: IMGCHHorizontal direction brightness correction image;IMGCVVertical direction brightness correction image;IMGCorrect—
Brightness correction image.
Described step 6) in, first with step 5) the brightness correction image that obtains carries out binaryzation by segmentation threshold, with 1
Representing defect, 0 represents background, it is thus achieved that defect binary image;Again defect binary image is iterated by below equation
Hole is filled out in process, then carries out 4 × 4 medium filterings, obtains surface defect image:
Fk=(Fk-1E)?Dc (8)
In formula, F processing result image;The data of D defect binary image, DcThe supplementary set of D;E four connected region;
K calculation times, k=1,2,3 ..., work as k=1, F0It it is binary image.
The invention have the advantages that:
The present invention utilizes the gray scale discontinuity feature of fruit surface defect image-region, by method of least square quadratic term
Multinomial completes brightness of image and returns with image adaptive brightness correction without manual debugging setting, it is only necessary to one simple complete
Office's threshold value completes defect Segmentation and can effectively detect surface defect.Algorithm simplicity program engineering realizes, at fruit and agricultural production
Product quality computers vision on-line checking aspect has bigger application potential.
Accompanying drawing explanation
Fig. 1 is the flow chart of the inventive method.
Fig. 2 is original color image in the embodiment of the present invention 1.
Fig. 3 is fruit gray-scale map in the embodiment of the present invention 1.
Fig. 4 is horizontal direction regression figure picture in the embodiment of the present invention 1.
Fig. 5 is vertical direction regression figure picture in the embodiment of the present invention 1.
Fig. 6 be in the embodiment of the present invention 1 level to brightness correction image.
Fig. 7 is vertical direction brightness correction image in the embodiment of the present invention 1.
Fig. 8 is brightness correction image in the embodiment of the present invention 1.
Fig. 9 is binary image in the embodiment of the present invention 1.
Figure 10 is four connected region in the embodiment of the present invention 1.
Figure 11 is extraction surface defect image in the embodiment of the present invention 1.
Figure 12 is Citrus sinensis Osbeck wind wound fruit original graph.
Figure 13 is Citrus sinensis Osbeck wind wound fruit surface defects detection result figure.
Figure 14 is Citrus sinensis Osbeck thrips fruit original graph.
Figure 15 is Citrus sinensis Osbeck thrips fruit surface defects detection result figure.
Figure 16 is Citrus sinensis Osbeck shell wormed fruit original graph.
Figure 17 is Citrus sinensis Osbeck shell wormed fruit surface defects detection result figure.
Figure 18 is Citrus sinensis Osbeck ulcer fruit original graph.
Figure 19 is Citrus sinensis Osbeck ulcer fruit surface defects detection result figure.
Figure 20 is Citrus sinensis Osbeck dehiscent fruit original graph.
Figure 21 is Citrus sinensis Osbeck dehiscent fruit surface defects detection result figure.
Detailed description of the invention
The present invention is further illustrated with specific embodiment below in conjunction with the accompanying drawings.
As it is shown in figure 1, the present embodiment is as follows:
1) shooting fruit RGB color image, as shown in Figure 2;
2) fruit RGB color image is converted to initial gray image
3) carry out binaryzation after extracting fruit object removal background, obtain binaryzation fruit image, represent fruit with 1,0 table
Show background;
4) by step 3) the binaryzation fruit image and the step 2 that obtain) gray level image that obtains carries out individual element point phase
Take advantage of, obtain fruit gray level image as shown in Figure 3, fruit gray level image is carried out twice duplication, respectively horizontal direction regression figure
Picture and vertical direction regression figure picture.
5) start scanning at fruit gray level image top line, find the 1st row gray value pixel quantity more than 0 big
In the row of 3, by the gray value of this row pixel, by order write vector f (n) from left to right, (n is the row that grey scale pixel value is corresponding
Number).
6) by formula (1), f (n) is carried out least square regression, obtain the coefficient k in formula (1)2、k1、k0。
7) in formula (2), poor ε is taken0It is 0, brings k into2、k1、k0, row n calculate and return gray value, and write level side
To regression figure as corresponding line, column position.
8) step 5 is repeated)-7), until the bottom line of fruit gray level image.Obtain horizontal direction as shown in Figure 4
Regression figure picture.
9) start scanning at fruit gray level image Far Left string, find the 1st row gray value pixel quantity more than 0 big
In the row of 3, by the gray value of this row pixel, by order write vector f (n) from top to bottom, (n is the row that grey scale pixel value is corresponding
Number).
10) by formula (3), g (m) is carried out least square regression, obtain the coefficient p in formula (3)2、p1、p0。
11) in formula (4), poor ε is taken1It is 0, brings p into2、p1、p0, line number m calculates and returns gray value, and write vertical
Direction regression figure is as corresponding line, column position.
12) step 9 is repeated)-11), until the rightmost string of fruit gray level image.Obtain Vertical Square as shown in Figure 5
To regression figure picture.
13) use formula (5) to obtain horizontal direction brightness correction image as shown in Figure 6, use formula (6) to obtain such as figure
Vertical direction brightness correction image shown in 7, uses formula (7) to obtain brightness correction image as shown in Figure 8.
14) utilize step 13) in brightness correction image carry out binaryzation, represent defect with 1,0 represent background, it is thus achieved that as
Binary image shown in Fig. 9, binary image is iterated processing by formula (8), then uses four as shown in Figure 10
Connected domain carries out 4 × 4 medium filterings, obtains surface defect image as shown in figure 11, and wherein white portion is defect area.
The present invention is successively to Citrus sinensis Osbeck wind wound fruit, Citrus sinensis Osbeck thrips fruit, Citrus sinensis Osbeck shell wormed fruit, Citrus sinensis Osbeck ulcer fruit and Citrus sinensis Osbeck dehiscent fruit
Fruit image detects, and original case and the testing result of Citrus sinensis Osbeck wind wound fruit are distinguished the most as shown in Figure 12 and Figure 13, Citrus sinensis Osbeck thrips
The most as shown in Figure 14 and Figure 15, original case and the testing result of Citrus sinensis Osbeck shell wormed fruit are divided for the original case of fruit and testing result
The most as shown in Figure 16 and Figure 17, the most as shown in Figure 18 and Figure 19, Citrus sinensis Osbeck splits for the original case of Citrus sinensis Osbeck ulcer fruit and testing result
The original case of fruit and testing result are the most as shown in Figure 20 and Figure 21.Repeatedly implement through embodiment, the inventive method accurate
Rate can reach 95%.
Above-mentioned detailed description of the invention is used for illustrating the present invention rather than limiting the invention, the present invention's
In spirit and scope of the claims, any modifications and changes that the present invention is made, both fall within the protection model of the present invention
Enclose.
Claims (4)
1. one kind returns the fruit surface defect detection method corrected based on adaption brightness, it is characterised in that the method includes
Following steps:
1) obtain fruit RGB color image, fruit RGB color image is converted to initial gray image;
2) by initial gray image segmentation background and target, carry out binary conversion treatment and obtain binaryzation fruit image, represent water with 1
Really target, 0 represents background;
3) by step 2) the binaryzation fruit image and the step 1 that obtain) the initial gray image that obtains carries out individual element point phase
Take advantage of, obtain fruit gray level image;
4) all pixels that every a line in fruit gray level image and every string remove background pixel carry out method of least square respectively
The polynomial regression Calculation of quadratic term, obtains horizontal direction regression figure picture and vertical direction regression figure picture respectively;
5) fruit gray level image is respectively divided by horizontal direction regression figure picture and vertical direction regression figure picture acquisition horizontal direction is bright
Degree correcting image and vertical direction brightness correction image, by horizontal direction brightness correction image and vertical direction brightness correction image
Addition is averaged acquisition brightness correction image;
6) by segmentation threshold, brightness correction image is carried out binaryzation, then complete to fill out hole and medium filtering obtains fruit surface
Defect image.
A kind of fruit surface defect detection method returning rectification based on adaption brightness the most according to claim 1, its
It is characterised by: described step 4) particularly as follows:
4.1) start to scan successively each more than 3 of the gray value pixel quantity more than 0 at fruit gray level image top line
OK, for every a line, by the gray value of this row gray value pixel more than 0 by from left to right order writing line vector f successively
(n);
4.2) as follows row vector f (n) is carried out least square regression calculating, obtain row quadratic term coefficient k2, row first order
Coefficient k1With row constant term k0;
f1(n)=k2n2+k1n+k0+ε0
Wherein, f1(n) grey scale pixel value;The row number at n grey scale pixel value place;k2Row quadratic term coefficient;k1Row is once
Term coefficient;k0Row constant term;ε0Row residual error;
4.3) row residual epsilon is taken again0It is 0, by row quadratic term coefficient k2, row Monomial coefficient k1With row constant term k0, grey scale pixel value
The row n at place substitutes into below equation and calculates acquisition row recurrence gray value f again2(n):
f2(n)=k2n2+k1n+k0+ε0
Wherein, f2N () returns gray value for row;
4.4) using calculating acquisition row recurrence gray value as the pixel value of this location of pixels in horizontal direction regression figure picture, for ash
The angle value pixel quantity more than 0 retains original gray value as pixel value less than in the row horizontal direction regression figure picture of 3;
4.5) start to scan successively each more than 3 of the gray value pixel quantity more than 0 at fruit gray level image Far Left string
Row, for every string, write column vector g by the gray value of this row gray value pixel more than 0 by order from top to bottom successively
(m);
4.6) as follows column vector g (m) is carried out least square regression calculating, obtain row quadratic term coefficient p2, row first order
Coefficient p1With row constant term p0;
g1(m)=p2m2+p1m+p0+ε1
Wherein, g1(m) grey scale pixel value;The line number at m grey scale pixel value place;p2Row quadratic term coefficient;p1Row are once
Term coefficient;p0Row constant term;ε1Row residual error;
4.7) row residual epsilon is taken again1It is 0, by row quadratic term coefficient p2, row Monomial coefficient p1With row constant term p0, grey scale pixel value
The row m at place substitutes into below equation and calculates acquisition row recurrence gray value g again2(m):
g2(m)=p2m2+p1m+p0+ε1
Wherein, g2M () returns gray value for row;
4.8) using calculating acquisition row recurrence gray value as the pixel value of this location of pixels in vertical direction regression figure picture, for ash
The angle value pixel quantity more than 0 retains original gray value as pixel value less than in the row vertical direction regression figure picture of 3.
A kind of fruit surface defect detection method returning rectification based on adaption brightness the most according to claim 1, its
It is characterised by: described step 5) particularly as follows:
5.1) below equation is used to obtain horizontal direction brightness correction image,
IMGCH=IMGFG/IMGRH
In formula: IMGCHHorizontal direction brightness correction image;IMGRHHorizontal direction regression figure picture;IMGFHFruit gray-scale map
Picture;
5.2) below equation is used to obtain vertical direction brightness correction image,
IMGCV=IMGFG/IMGRV
In formula: IMGCVVertical direction brightness correction image;IMGRVVertical direction regression figure picture;IMGFGFruit gray-scale map
Picture;
5.3) use below equation acquisition brightness correction image:
IMGCorrect=0.5 × MGCH+0.5×IMGCV
In formula: IMGCHHorizontal direction brightness correction image;IMGCVVertical direction brightness correction image;IMGCorrectBrightness
Correcting image.
A kind of fruit surface defect detection method returning rectification based on adaption brightness the most according to claim 1, its
It is characterised by: described step 6) in, first with step 5) the brightness correction image that obtains carries out binaryzation by segmentation threshold, uses
1 represents defect, and 0 represents background, it is thus achieved that defect binary image;
Again defect binary image is iterated process by below equation and fills out hole, then carry out 4 × 4 medium filterings, obtain
Surface defect image:
Fk=(Fk-1E)?Dc
In formula, F processing result image;The data of D defect binary image, DcThe supplementary set of D;E four connected region;K counts
Calculation number of times, k=1,2,3 ....
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610349350.8A CN106124511B (en) | 2016-05-24 | 2016-05-24 | A kind of fruit surface defect detection method returning correction based on adaption brightness |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610349350.8A CN106124511B (en) | 2016-05-24 | 2016-05-24 | A kind of fruit surface defect detection method returning correction based on adaption brightness |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106124511A true CN106124511A (en) | 2016-11-16 |
CN106124511B CN106124511B (en) | 2018-11-06 |
Family
ID=57270016
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610349350.8A Active CN106124511B (en) | 2016-05-24 | 2016-05-24 | A kind of fruit surface defect detection method returning correction based on adaption brightness |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106124511B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107037056A (en) * | 2017-03-29 | 2017-08-11 | 浙江大学 | A kind of fruit surface defect detection method of quick Local threshold segmentation |
CN108596909A (en) * | 2018-05-09 | 2018-09-28 | 芜湖哈特机器人产业技术研究院有限公司 | Hollow cylinder Surface Flaw Detection system and method |
CN109060842A (en) * | 2018-08-14 | 2018-12-21 | 西南大学 | Citrus detection method of surface flaw, computer program based on surface fitting correction |
CN109613023A (en) * | 2019-01-28 | 2019-04-12 | 西南大学 | A kind of fruit surface defect rapid detection method of regional luminance adaptively correcting |
CN110060249A (en) * | 2019-04-22 | 2019-07-26 | 淮海工学院 | A kind of detection method applying mud wood moulding surface small point-like defect |
CN114354628A (en) * | 2022-01-05 | 2022-04-15 | 威海若维信息科技有限公司 | Root agricultural product defect detection method based on machine vision |
CN116645364A (en) * | 2023-07-18 | 2023-08-25 | 金乡县金沪合金钢有限公司 | Alloy steel casting air hole defect detection method based on image data |
CN117079144A (en) * | 2023-10-17 | 2023-11-17 | 深圳市城市交通规划设计研究中心股份有限公司 | Linear crack extraction method for asphalt pavement detection image under non-uniform illumination |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102788806A (en) * | 2012-07-19 | 2012-11-21 | 北京农业智能装备技术研究中心 | Fruit peel defect detection method based on spheroidic brightness transformation |
JP2012230081A (en) * | 2011-04-27 | 2012-11-22 | Panasonic Corp | External appearance inspection device of glossy product and program |
CN104677920A (en) * | 2015-02-09 | 2015-06-03 | 浙江大学 | Fruit surface bug detection marking method based on local grey scale mutation on fruit surface |
CN105004737A (en) * | 2015-07-14 | 2015-10-28 | 浙江大学 | Self-adaption improved gradient information-based fruit surface defect detection method |
CN105044128A (en) * | 2015-07-14 | 2015-11-11 | 浙江大学 | Method for detecting fruit surface defect based on light background difference image |
CN105354847A (en) * | 2015-11-10 | 2016-02-24 | 浙江大学 | Fruit surface defect detection method based on adaptive segmentation of sliding comparison window |
-
2016
- 2016-05-24 CN CN201610349350.8A patent/CN106124511B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012230081A (en) * | 2011-04-27 | 2012-11-22 | Panasonic Corp | External appearance inspection device of glossy product and program |
CN102788806A (en) * | 2012-07-19 | 2012-11-21 | 北京农业智能装备技术研究中心 | Fruit peel defect detection method based on spheroidic brightness transformation |
CN104677920A (en) * | 2015-02-09 | 2015-06-03 | 浙江大学 | Fruit surface bug detection marking method based on local grey scale mutation on fruit surface |
CN105004737A (en) * | 2015-07-14 | 2015-10-28 | 浙江大学 | Self-adaption improved gradient information-based fruit surface defect detection method |
CN105044128A (en) * | 2015-07-14 | 2015-11-11 | 浙江大学 | Method for detecting fruit surface defect based on light background difference image |
CN105354847A (en) * | 2015-11-10 | 2016-02-24 | 浙江大学 | Fruit surface defect detection method based on adaptive segmentation of sliding comparison window |
Non-Patent Citations (2)
Title |
---|
JIANGBO LI ET AL.: "Automatic detection of common surface defects on oranges using combined lighting transform and image ratio methods", 《POSTHARVEST BIOLOGY AND TECHNOLOGY》 * |
王福杰等: "苹果图像的背景分割与目标提取", 《农业机械学报》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107037056A (en) * | 2017-03-29 | 2017-08-11 | 浙江大学 | A kind of fruit surface defect detection method of quick Local threshold segmentation |
CN108596909A (en) * | 2018-05-09 | 2018-09-28 | 芜湖哈特机器人产业技术研究院有限公司 | Hollow cylinder Surface Flaw Detection system and method |
CN109060842A (en) * | 2018-08-14 | 2018-12-21 | 西南大学 | Citrus detection method of surface flaw, computer program based on surface fitting correction |
CN109613023A (en) * | 2019-01-28 | 2019-04-12 | 西南大学 | A kind of fruit surface defect rapid detection method of regional luminance adaptively correcting |
CN109613023B (en) * | 2019-01-28 | 2021-02-26 | 西南大学 | Fruit surface defect rapid detection method with area brightness self-adaptive correction |
CN110060249A (en) * | 2019-04-22 | 2019-07-26 | 淮海工学院 | A kind of detection method applying mud wood moulding surface small point-like defect |
CN114354628A (en) * | 2022-01-05 | 2022-04-15 | 威海若维信息科技有限公司 | Root agricultural product defect detection method based on machine vision |
CN114354628B (en) * | 2022-01-05 | 2023-10-27 | 威海若维信息科技有限公司 | Rhizome agricultural product defect detection method based on machine vision |
CN116645364A (en) * | 2023-07-18 | 2023-08-25 | 金乡县金沪合金钢有限公司 | Alloy steel casting air hole defect detection method based on image data |
CN116645364B (en) * | 2023-07-18 | 2023-10-27 | 金乡县金沪合金钢有限公司 | Alloy steel casting air hole defect detection method based on image data |
CN117079144A (en) * | 2023-10-17 | 2023-11-17 | 深圳市城市交通规划设计研究中心股份有限公司 | Linear crack extraction method for asphalt pavement detection image under non-uniform illumination |
CN117079144B (en) * | 2023-10-17 | 2023-12-26 | 深圳市城市交通规划设计研究中心股份有限公司 | Linear crack extraction method for asphalt pavement detection image under non-uniform illumination |
Also Published As
Publication number | Publication date |
---|---|
CN106124511B (en) | 2018-11-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106124511A (en) | A kind of fruit surface defect detection method returning rectification based on adaption brightness | |
CN105675625B (en) | A kind of fruit surface defect detection method of Gradient Iteration Threshold segmentation | |
CN105354847B (en) | Fruit surface defect detection method based on sliding comparison window adaptivenon-uniform sampling | |
CN105004737B (en) | Fruit surface defect detection method based on adaptive impovement type gradient information | |
CN105354865B (en) | The automatic cloud detection method of optic of multispectral remote sensing satellite image and system | |
CN105044128B (en) | Fruit surface defect detection method based on illumination background difference shadow method | |
CN109613023A (en) | A kind of fruit surface defect rapid detection method of regional luminance adaptively correcting | |
CN111383209A (en) | Unsupervised flaw detection method based on full convolution self-encoder network | |
CN105741276B (en) | A kind of ship waterline extracting method | |
CN109060842B (en) | Citrus surface defect detection method based on surface fitting correction | |
Jin et al. | Perceptual image quality assessment using block-based multi-metric fusion (BMMF) | |
CN103164855A (en) | Bayesian Decision Theory foreground extraction method combined with reflected illumination | |
CN108416803A (en) | A kind of scene depth restoration methods of the Multi-information acquisition based on deep neural network | |
CN109064479B (en) | Sea-sky-line detection method based on gray dynamic features of adjacent video frames | |
CN110853027A (en) | Three-dimensional synthetic image no-reference quality evaluation method based on local variation and global variation | |
CN107037056A (en) | A kind of fruit surface defect detection method of quick Local threshold segmentation | |
CN106157264A (en) | Large area image uneven illumination bearing calibration based on empirical mode decomposition | |
CN108305232A (en) | A kind of single frames high dynamic range images generation method | |
CN110458812A (en) | A kind of similar round fruit defects detection method based on color description and sparse expression | |
CN114266768A (en) | Method for generating surface scratch defect image | |
CN109978916A (en) | Vibe moving target detecting method based on gray level image characteristic matching | |
CN105528772B (en) | A kind of image interfusion method based on directiveness filtering | |
CN102629369B (en) | Single color image shadow removal method based on illumination surface modeling | |
CN116500052A (en) | Edible oil impurity visual detection system and application method thereof | |
CN116402802A (en) | Underwater image quality evaluation method based on color space multi-feature fusion |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20210527 Address after: 310012 room 1102, block B, Lishui digital building, 153 Lianchuang street, Wuchang Street, Yuhang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou nuotian Intelligent Technology Co.,Ltd. Address before: 310027 No. 38, Zhejiang Road, Hangzhou, Zhejiang, Xihu District Patentee before: ZHEJIANG University |
|
TR01 | Transfer of patent right |