CN107590801B - Image processing method for beef eye muscle marbling segmentation and grading - Google Patents
Image processing method for beef eye muscle marbling segmentation and grading Download PDFInfo
- Publication number
- CN107590801B CN107590801B CN201710803346.9A CN201710803346A CN107590801B CN 107590801 B CN107590801 B CN 107590801B CN 201710803346 A CN201710803346 A CN 201710803346A CN 107590801 B CN107590801 B CN 107590801B
- Authority
- CN
- China
- Prior art keywords
- image
- eye muscle
- beef
- region
- marbling
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides an image processing method for beef eye muscle marbling segmentation and grading. The image processing method provided by the invention specifically comprises the following steps: segmenting an eye muscle section image of the cattle carcass; and (4) extracting marbling in the beef eye muscle area, and grading the marbling of the beef eye muscles. The invention relates to a method for automatically segmenting the eye muscles of beef, extracting marbling in the eye muscles and objectively and efficiently grading the marbling based on computer vision and image processing technology. The invention is suitable for but not limited to beef production enterprises and quality detection departments, and has good market application prospect.
Description
Technical Field
The invention relates to a computer vision and image processing technology, in particular to an image processing method for beef eye muscle marbling segmentation and grading.
Background
Beef marbling is an important criterion for beef grade assessment. At present, the beef grading method of all countries in the world generally adopts subjective artificial visual evaluation. A beef grader evaluates the marbling grade of beef by observing the richness of intramuscular fat (marbling) at the transverse section of the longest muscle (eye muscle) between the thoraco-costal back of a beef carcass, and finally evaluates the quality grade of the beef by referring to the flesh color, the fat color and the physiological maturity of the beef. The advantage of the manual visual rating method is that the grader's intuition and experience can ensure a high beef rating accuracy, and this rating process is lossless. However, the artificial visual rating method still has some fundamental disadvantages: firstly, the method is subjective, and when two graders grade the same beef, the evaluation results may be different; secondly, inconsistency exists, and when the same grader grades the same beef twice, the evaluation results of the beef may be different; thirdly, the method has high labor cost and low efficiency in rating.
Disclosure of Invention
The invention aims to provide an image processing method for beef eye muscle marbling segmentation and grading. The invention solves the problems of subjectivity, inconsistency, high rating cost, low efficiency and the like of the manual rating method.
In order to realize the purpose of the invention, the invention adopts the following technical scheme to realize:
the invention provides an image processing method for beef eye muscle marbling segmentation and grading, which comprises the following steps:
1) segmenting an eye muscle section image of the beef;
(1) performing image graying on the beef eye muscle section color image shot by a digital camera to obtain an eye muscle section grayed image;
(2) processing the eye muscle section gray image to obtain an eye muscle section binary image;
(3) carrying out connected region marking on the eye muscle section binary image;
(4) searching a connected region mark corresponding to the central point region of the eye muscle section grayed image, setting corresponding eye muscle section grayed image pixels under the marked region as a foreground, and resetting all pixels of the rest marked regions as a background to obtain an eye muscle initial region image;
(5) setting a foreground of the eye muscle initial region image as one, and clearing a background to obtain a binary image of the eye muscle initial region;
(6) calculating a convex hull of the eye muscle initial region binary image;
(7) performing an active contour segmentation algorithm on the eye muscle section gray image by taking a convex hull boundary as an initial contour of an eye muscle region to obtain a beef eye muscle region contour;
2) extracting marbling in the beef eye muscle area;
(1) for the eye muscle gray image, reserving pixels contained in the outline of the beef eye muscle area, setting the pixels as a foreground area, and resetting the pixels outside the outline to be a background to obtain the beef eye muscle area image;
(2) processing the beef eye muscle region image to obtain a marbling binary image of the beef eye muscle region, wherein the foreground part of the binary image in the eye muscle range is identified as marbling, and the rest pixels in the eye muscle range are identified as eye muscle red meat;
(3) performing a superpixel segmentation algorithm on the beef eye muscle region image to obtain a superpixel grid block;
(4) taking out the super-pixel grid blocks of the foreground part of the eye muscle region image to obtain the eye muscle region super-pixel grid blocks;
(5) tiling the eye muscle region superpixel grid blocks on the marbling binary image, and calculating the fat content of each grid block; when the fat content in a certain super-pixel grid block exceeds a critical value, setting all pixels in the super-pixel grid block to be 1, determining that the super-pixel grid block is a large fat block, otherwise, setting all pixels in the super-pixel grid block to be 0, and obtaining a large fat block binary image in an eye muscle area;
3) beef eye marbling rating;
(1) calculating a fat ratio P, a large fat block ratio Q and a box-counting dimension Db corresponding to the sample image;
(2) establishing a following multiple linear regression equation set according to the corresponding relation between the marble pattern grade MB of the sample image and the fat proportion P, the large fat block proportion Q and the box-counting dimension Db;
in the formula: MB (multimedia broadcasting)iFor the marbleizing grade, P, corresponding to the ith sample imagei、Qi、DbiRespectively the fat ratio P, the large fat block ratio Q and the box-counting dimension Db, β calculated from the ith sample image0Is a constant term, β1、β2、β3Is a regression coefficient, εiThe regression error of the ith sample image;
(3) establishing a multivariate linear regression equation of beef marbling grade MB through a least square method, wherein the equation is shown as the following formula;
MB＝β0+β1P+β2Q+β3Db
in the formula: MB is the prediction grade of the marble patterns of the beef of the sample image, P is the fat proportion of the sample image, Q is the large fat block proportion of the sample image, and Db is the box-counting dimension of the sample image.
The invention has the advantages and technical effects that: the invention relates to a method for automatically segmenting the eye muscles of beef, extracting marbling in the eye muscles and objectively and efficiently grading the marbling based on computer vision and image processing technology. The invention is suitable for but not limited to beef production enterprises and quality detection departments, and has good market application prospect.
Drawings
FIG. 1 is a flow chart of an image processing method according to the present invention;
FIG. 2 is a gray scale image of a section of an eye muscle according to the present invention;
FIG. 3 is a binary image of the section of the eye muscle according to the present invention;
FIG. 4 is an image of an initial region of the eye muscles according to the present invention;
FIG. 5 is a binary image of an initial region of eye muscles according to the present invention;
FIG. 6 is a convex hull of a binary image of an initial region of eye muscles according to the present invention;
FIG. 7 is an initial contour of the eye muscle region of the present invention;
FIG. 8 is a contour of an eye muscle region according to the present invention;
FIG. 9 is an image of an eye muscle region according to the present invention;
fig. 10 is a marbleized binary image according to the present invention;
FIG. 11 is a superpixel gridblock in accordance with the present invention;
FIG. 12 is a superpixel grid block of the eye muscle region of the present invention;
FIG. 13 is a marbleized binary image of a superpixel grid of the tiled superior eye muscle area of the present invention;
FIG. 14 is a binary image of large fat blocks in the eye muscle region according to the present invention;
FIG. 15 is a plot of a fit between actual and predicted values of marbling levels for beef according to the invention.
Detailed Description
The invention is described in detail with reference to the accompanying drawings and specific embodiments.
As shown in the flow chart of fig. 1, the beef marbling grading method of the invention comprises the following steps:
firstly, beef eye muscle section image segmentation;
secondly, marbling extraction is carried out on the beef eye muscle area;
and (III) beef eye muscle marbling rating.
Example 1
This example describes the specific steps in detail.
Beef eye muscle section image segmentation
The method specifically comprises the following steps:
1. and (3) graying the image of the beef eye muscle section color image shot by the digital camera to obtain a grayed image of the eye muscle section (as shown in figure 2).
2. Processing the eye muscle section gray scale image by Otsu to obtain an eye muscle section binary image (as shown in FIG. 3).
3. And marking a connected region of the eye muscle section binary image.
4. And searching a connected region mark corresponding to the central point region of the eye muscle section grayed image, setting the corresponding eye muscle section grayed image pixel under the marked region as a foreground, and clearing all the pixels of the rest marked regions as a background to obtain an eye muscle initial region image (as shown in fig. 4).
5. And setting the foreground of the eye muscle initial region image as one, and clearing the background to zero to obtain a two-value image of the eye muscle initial region (as shown in figure 5).
6. And calculating a convex hull of the eye muscle initial region binary image (as shown in figure 6).
7. And (3) carrying out an active contour segmentation algorithm on the eye muscle section gray image by taking the convex hull boundary as an initial contour of an eye muscle region (as shown in fig. 7) to obtain a beef eye muscle region contour (as shown in fig. 8).
The active contour segmentation algorithm comprises the following steps:
(1) establishing an energy functional related to the image and the contour
Where I is the image, C (q) is the contour, q is a parameter of the contour, g is a decreasing function of the form:
where p is 2 and G is a gaussian filter function of the form:
where σ is a standard deviation parameter, and σ is selected to be 1.
(2) And carrying out a variation method on the established energy functional to obtain a profile time-varying equation in the following form:
where k is the curvature of the curve and N is the normal vector pointing into the interior of the profile curve.
(3) And introducing a level set u which changes along with time t to represent the contour, namely, if u (x, y, t) ═ 0 is the contour C, u (x, y, t) >0 is the contour containing region, and u (x, y, t) <0 is the region outside the contour, the contour can be represented as follows along with the time-varying equation in the form of the level set:
(4) and calculating a final contour curve by using a finite difference method for a contour time change equation and carrying out iteration (100 times in the method) on the time t for a given number of times.
Extraction of marbling in the region of the eye muscles
The method comprises the following specific steps:
1. for the eye muscle grayed image, the beef eye muscle region contour including pixels is reserved and set as a foreground region, and the pixels outside the contour are cleared and set as a background, so that the beef eye muscle region image is obtained (as shown in fig. 9).
2. The beef eye muscle region image is processed by the Otsu method (Otsu) to obtain a marbled binary image (as shown in FIG. 10) of the beef eye muscle region, wherein the foreground part of the binary image in the eye muscle region is identified as beef eye muscle fat (i.e. marbleizing), and the remaining pixels in the eye muscle region are identified as eye muscle red.
3. And (3) performing a superpixel segmentation algorithm on the beef eye muscle region image to obtain a superpixel grid block (as shown in fig. 11).
4. And (3) extracting the super-pixel grid blocks of the foreground part of the eye muscle region image to obtain the eye muscle region super-pixel grid blocks (as shown in figure 12).
5. And (3) tiling the eye muscle region superpixel grid blocks on the marbling binary image (as shown in figure 13), and calculating the fat content of each grid block. When the fat content in a super-pixel grid block exceeds a threshold value (0.3-0.5 can be selected in the invention, and 0.4 in the embodiment), all pixels in the super-pixel grid block are set to be 1, the super-pixel grid block is determined to be a large fat block, otherwise, all pixels in the super-pixel grid block are set to be 0, and a large fat block binary image of the eye muscle region is obtained (as shown in fig. 14).
The super-pixel segmentation algorithm in the step 3 comprises the following steps:
(1) determining the number k of superpixel blocks, wherein the initial side length S of each superpixel block is
Where N is the total number of pixels in the image.
(2) And numbering each super pixel center from 1 to k, the number of pixels belonging to the super pixel block being the same as the number of the super pixel centers.
(3) Finding out the center of the super pixel with the nearest generalized distance (including space and color) for each pixel of the image by using a k-means clustering method, wherein the generalized distance D is defined as
Where l, a, and b are CIELAB color space, x and y are coordinate space, and Nc and Ns are weight constants (in this method, Nc is 0.5, and Ns is 10).
(4) And changing the number of all pixels of the image into the number of the center of the super pixel closest to the generalized distance.
(5) And setting the gravity center of all the pixels belonging to the same new number as the new center of the numbered superpixel.
(6) And returning to the step 3, and obtaining the final superpixel number after a given number of iterations (100 times in the embodiment).
(III) beef Ocular Marble Pattern rating
Rating standard:
(1) fat ratio P
P is F/T type (8)
In the formula, F is the number of the marbleized pixel points in the marbleized binary image, and T is the number of the foreground pixel points in the beef eye muscle region image.
(2) Large fat mass fraction Q
Q ═ Fk/Tk formula (9)
In the formula, Fk is the number of large fat blocks in the large fat block binary image in the eye muscle region, and Tk is the number of all super pixels in the large fat block binary image in the eye muscle region.
(3) Box counting dimension Db
In the formula, s is the average side length of the superpixels, and N is the number of the superpixels of the large fat block in the large fat block binary image of the eye muscle region.
Specifically, the marble grading step:
1. and (3) calculating the fat ratio P, the large fat block ratio Q and the box-counting dimension Db corresponding to the sample images with different beef marbling grades according to the steps (I) and (II).
2. According to the corresponding relation between the marbling grade MB and the fat ratio P of different sample images, the large fat block ratio Q and the box-counting dimension Db, the following multiple linear regression equation set is established.
In the formula, MBiFor the marbleizing grade, P, corresponding to the ith sample imagei、Qi、DbiRespectively calculating the fat ratio P, the large fat block ratio Q and the counting box from the ith sample imageDimension Db, β0Is a constant term, β1、β2、β3Is a regression coefficient, εiThe regression error for the ith sample image.
3. And finally, establishing a multivariate linear regression equation of beef marbling grading by a least square method, wherein the equation is shown in the following formula.
MB＝β0+β1P+β2Q+β3Db type (12)
In the formula, MB is the prediction grade of the beef marbling of the sample image, P is the fat proportion of the sample image, Q is the large fat block proportion of the sample image, and Db is the box-counting dimension of the sample image.
Example 2
The steps (a) and (b) in example 1 were repeated for 18 sample images having a typical beef marbling level to obtain correspondence data of the beef marbling level (MB) to the ratio of eye muscle fat (P), the ratio of eye muscle large fat masses (Q), the box-counting dimension (Db), and the information dimension (Di) for 18 groups of sample images, as shown in table 1:
table 118 sets of correspondence data of MB, P, Q, Db, Di of sample images
According to the data of the corresponding relation between the marble pattern grade (MB) of the sample image and the ratio of eye muscle fat (P), the proportion of large fat blocks of the eye muscle (Q) and the box-counting dimension (Db), the following equation system can be established:
determining the regression coefficients β by least squares0、β1、β2、β3. Finally establishing a beef marbling grade multi-element linear regression equationThe following were used:
MB＝5.820P+12.496Q-0.385Db+0.0527
according to the multivariate linear regression equation of the beef marbling grade established in the step three, a fitting curve (shown in figure 15) between the actual value and the predicted value of the beef marbling grade (MB) can be obtained, the fitting degree is 0.9654, and the multivariate linear regression equation model established by the invention can effectively predict the beef marbling grade.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing examples, it will be apparent to those skilled in the art that various changes in the embodiments and modifications can be made, and equivalents can be substituted for elements thereof; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions.
Claims (6)
1. An image processing method for beef eye muscle marbling segmentation and grading, characterized in that it comprises the following steps:
1) segmenting an eye muscle section image of the beef;
(1) performing image graying on the beef eye muscle section color image shot by a digital camera to obtain an eye muscle section grayed image;
(2) processing the eye muscle section gray image to obtain an eye muscle section binary image;
(3) carrying out connected region marking on the eye muscle section binary image;
(4) searching a connected region mark corresponding to the central point region of the eye muscle section grayed image, setting corresponding eye muscle section grayed image pixels under the marked region as a foreground, and resetting all pixels of the rest marked regions as a background to obtain an eye muscle initial region image;
(5) setting a foreground of the eye muscle initial region image as one, and clearing a background to obtain a binary image of the eye muscle initial region;
(6) calculating a convex hull of the eye muscle initial region binary image;
(7) performing an active contour segmentation algorithm on the eye muscle section gray image by taking a convex hull boundary as an initial contour of an eye muscle region to obtain a beef eye muscle region contour;
2) extracting marbling in the beef eye muscle area;
(1) for the eye muscle gray image, reserving pixels contained in the outline of the beef eye muscle area, setting the pixels as a foreground area, and resetting the pixels outside the outline to be a background to obtain the beef eye muscle area image;
(2) processing the beef eye muscle region image to obtain a marbling binary image of the beef eye muscle region, wherein the foreground part of the binary image in the eye muscle range is identified as marbling, and the rest pixels in the eye muscle range are identified as eye muscle red meat;
(3) performing a superpixel segmentation algorithm on the beef eye muscle region image to obtain a superpixel grid block;
(4) taking out the super-pixel grid blocks of the foreground part of the eye muscle region image to obtain the eye muscle region super-pixel grid blocks;
(5) tiling the eye muscle region superpixel grid blocks on the marbling binary image, and calculating the fat content of each grid block; when the fat content in a certain super-pixel grid block exceeds a critical value, setting all pixels in the super-pixel grid block to be 1, determining that the super-pixel grid block is a large fat block, otherwise, setting all pixels in the super-pixel grid block to be 0, and obtaining a large fat block binary image in an eye muscle area;
3) beef eye marbling rating;
(1) calculating a fat ratio P, a large fat block ratio Q and a box-counting dimension Db corresponding to the sample image;
(2) establishing a following multiple linear regression equation set according to the corresponding relation between the marble pattern grade MB of the sample image and the fat proportion P, the large fat block proportion Q and the box-counting dimension Db;
in the formula: MB (multimedia broadcasting)iFor the marbleizing grade, P, corresponding to the ith sample imagei、Qi、DbiRespectively the fat ratio P, the large fat block ratio Q and the box-counting dimension Db, β calculated from the ith sample image0Is a constant term, β1、β2、β3Is a regression coefficient, εiThe regression error of the ith sample image;
(3) establishing a multivariate linear regression equation of beef marbling grade MB through a least square method, wherein the equation is shown as the following formula;
MB＝β0+β1P+β2Q+β3Db
in the formula: MB is the prediction grade of marble patterns of the beef of the sample image, P is the fat proportion of the sample image, Q is the large fat block proportion of the sample image, and Db is the box-counting dimension of the sample image;
the large fat block proportion Q is Fk/Tk, Fk is the number of large fat blocks in the binary image of the large fat block in the eye muscle region, and Tk is the number of all super pixels in the binary image of the large fat block in the eye muscle region;
the box-counting dimension
s is the average side length of the superpixels, and N is the number of the superpixels of the large fat block in the binary image of the large fat block in the eye muscle region.
2. The image processing method according to claim 1, characterized in that: the active contour segmentation algorithm in the step 1) comprises the following steps:
(1) establishing an energy functional with respect to an image and a contour
Where I is the image, C (q) is the contour, q is a parameter of the contour, g is a decreasing function of the form:
where p is 2 and G is a gaussian filter function of the form:
wherein sigma is a standard deviation parameter, and sigma is selected to be 1;
(2) and (3) carrying out a variation method on the established energy functional to obtain a profile time-varying equation in the following form:
wherein k is the curvature of the curve and N is a normal vector pointing to the inside of the profile curve;
(3) introducing a level set u which changes with time t to represent the contour, namely, if u (x, y, t) ═ 0 is the contour C, u (x, y, t) >0 is the contour-containing region, and u (x, y, t) <0 is the region outside the contour, then the contour time-varying equation is expressed in the form of a level set as follows:
(4) for the equation of the profile changing along with the time, a finite difference method is used for iterating the time t for a given number of times, and a final profile curve is calculated.
3. The image processing method according to claim 1, characterized in that: the super-pixel segmentation algorithm in the step 2) comprises the following steps:
(1) determining the number k of superpixel blocks, wherein the initial side length S of each superpixel block is
Wherein N is the total number of pixels of the image;
(2) numbering each superpixel center from 1 to k, wherein the pixel number belonging to the superpixel block is the same as the superpixel center number;
(3) finding out the center of the super pixel with the nearest generalized distance D defined as
Wherein l, a, b are CIELAB color space, x, y are coordinate space, Nc and Ns are weight constants (in this method, Nc is 0.5, Ns is 10);
(4) changing the serial numbers of all pixels of the image into the serial number of the center of the super pixel closest to the generalized distance;
(5) setting the gravity center of all pixels belonging to the same new number as the new center of the numbered superpixel;
(6) and returning to the step 3, and obtaining the final superpixel number through given times of iteration.
4. The image processing method according to claim 3, characterized in that: the iteration times in the step (6) are 100 times.
5. The image processing method according to claim 1, characterized in that: the fat ratio P is equal to F/T, F is the number of marbleized pixel points in the marbleized binary image, and T is the number of foreground pixel points in the beef eye muscle region image.
6. The image processing method according to claim 1, characterized in that: the critical value in step 2) (5) is 0.3-0.5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710803346.9A CN107590801B (en) | 2017-09-08 | 2017-09-08 | Image processing method for beef eye muscle marbling segmentation and grading |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710803346.9A CN107590801B (en) | 2017-09-08 | 2017-09-08 | Image processing method for beef eye muscle marbling segmentation and grading |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107590801A CN107590801A (en) | 2018-01-16 |
| CN107590801B true CN107590801B (en) | 2020-03-17 |
Family
ID=61050937
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710803346.9A Expired - Fee Related CN107590801B (en) | 2017-09-08 | 2017-09-08 | Image processing method for beef eye muscle marbling segmentation and grading |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107590801B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108535292B (en) * | 2018-03-28 | 2020-08-11 | 东南大学 | Method for determining upper limit of fine-grained soil mixing amount in high-speed railway base filler |
| CN110264464A (en) * | 2019-06-26 | 2019-09-20 | 西北农林科技大学 | A kind of beef cattle intramuscular fat content prediction model, its construction method and application |
| CN110555384A (en) * | 2019-07-31 | 2019-12-10 | 四川省草原科学研究院 | Beef marbling automatic grading system and method based on image data |
| CN116823857B (en) * | 2023-07-25 | 2024-03-19 | 查维斯机械制造(北京)有限公司 | Slaughter line pig carcass shearing intelligent positioning method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101561402A (en) * | 2009-05-07 | 2009-10-21 | 浙江大学 | Machine vision-based real-time detection and grading method and machine vision-based real-time detection and grading device for pork appearance quality |
| CN101706445A (en) * | 2009-11-10 | 2010-05-12 | 吉林大学 | Beef marbling grade scoring method and device |
| CN102156129A (en) * | 2009-12-02 | 2011-08-17 | 南京农业大学 | Beef quality intelligent grading system and method based on machine vision |
-
2017
- 2017-09-08 CN CN201710803346.9A patent/CN107590801B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101561402A (en) * | 2009-05-07 | 2009-10-21 | 浙江大学 | Machine vision-based real-time detection and grading method and machine vision-based real-time detection and grading device for pork appearance quality |
| CN101706445A (en) * | 2009-11-10 | 2010-05-12 | 吉林大学 | Beef marbling grade scoring method and device |
| CN102156129A (en) * | 2009-12-02 | 2011-08-17 | 南京农业大学 | Beef quality intelligent grading system and method based on machine vision |
Non-Patent Citations (6)
| Title |
|---|
| Geodesic Active Contours;Vicnet Caselles et al.;《International Journal of Computer Vision》;19971231;第22卷(第1期);第2.2-2.4节 * |
| Improved SLIC imagine segmentation algorithm based on K-means;Chun-yan Han;《Cluster Comput》;20170225;第2部分 * |
| 基于分形维和图像特征的牛肉大理石花纹等级判定模型;陈坤杰 等;《农业机械学报》;20120531;第43卷(第5期);第1-2部分 * |
| 基于该机K-Means的腹内脂肪自动定量检测算法;曹鸿吉 等;《计算机辅助设计与图形学学报》;20170430;第29卷(第4期);摘要 * |
| 牛肉图像中大理石花纹的提取技术研究;黄乐;《中国优秀硕士学位论文全文数据库 信息科技辑》;20111215;正文第三章 * |
| 牛胴体眼肌切面图像的分割研究;秦春芳;《中国优秀博硕士学位论文全文数据库（硕士） 工程科技I辑》;20070215;正文第三-四章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107590801A (en) | 2018-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107590801B (en) | Image processing method for beef eye muscle marbling segmentation and grading | |
| CN109191459B (en) | Automatic identification and rating method for continuous casting billet macrostructure center segregation defect | |
| CN106056118B (en) | A kind of identification method of counting for cell | |
| CN104933709B (en) | Random walk CT lung tissue image automatic segmentation methods based on prior information | |
| CN107392232B (en) | Flotation working condition classification method and system | |
| CN111582111B (en) | Cell component segmentation method based on semantic segmentation | |
| CN106340016B (en) | A kind of DNA quantitative analysis method based on microcytoscope image | |
| CN104792792A (en) | Stepwise-refinement pavement crack detection method | |
| CN112396619B (en) | Small particle segmentation method based on semantic segmentation and internally complex composition | |
| CN104268600B (en) | A kind of mineral flotation foam image texture analysis and operating mode's switch method based on Minkowski distances | |
| CN108681689B (en) | Frame rate enhanced gait recognition method and device based on generation of confrontation network | |
| CN113658117B (en) | Method for identifying and dividing aggregate boundary in asphalt mixture based on deep learning | |
| CN105160355A (en) | Remote sensing image change detection method based on region correlation and visual words | |
| CN110378924A (en) | Level set image segmentation method based on local entropy | |
| CN116342586B (en) | Road surface quality detection method based on machine vision | |
| CN112258525B (en) | Image abundance statistics and population identification algorithm based on bird high-frame frequency sequence | |
| CN115205289B (en) | Vision-based cork wood floor raw material grading method | |
| CN106327501A (en) | Quality evaluation method for thangka image with reference after repair | |
| CN109003287A (en) | Image partition method based on improved adaptive GA-IAGA | |
| CN108345898A (en) | A kind of novel line insulator Condition assessment of insulation method | |
| CN117853722A (en) | Steel metallographic structure segmentation method integrating superpixel information | |
| CN109978858B (en) | Double-frame thumbnail image quality evaluation method based on foreground detection | |
| CN109191482B (en) | Image merging and segmenting method based on regional adaptive spectral angle threshold | |
| CN117237595B (en) | Rapid extraction method for scratch area of intelligent watch shell | |
| CN109948708A (en) | Multispectral image feature level information fusion method when more based on the implicit canonical of iteration |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200317 Termination date: 20210908 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |