CN109584253B - Oil abrasive particle image segmentation method - Google Patents
Oil abrasive particle image segmentation method Download PDFInfo
- Publication number
- CN109584253B CN109584253B CN201811559431.6A CN201811559431A CN109584253B CN 109584253 B CN109584253 B CN 109584253B CN 201811559431 A CN201811559431 A CN 201811559431A CN 109584253 B CN109584253 B CN 109584253B
- Authority
- CN
- China
- Prior art keywords
- region
- image
- segmentation
- color
- watershed
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/155—Segmentation; Edge detection involving morphological operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses an oil abrasive particle image segmentation method, which comprises the following steps: 1) finishing watershed segmentation of the abrasive grain image based on a self-adaptive H-minima technology; 2) extracting color characteristic parameters of the abrasive particle image based on a Lab color space; 3) extracting texture characteristic parameters of the abrasive grain image based on the LBP map; 4) determining similarity of watershed homogeneous regions based on the Papanicolaou distance; 5) automatically marking seed regions based on the similarity matrix; 6) merging the watershed homogeneous regions based on a region merging criterion; 7) and correcting the region combination image based on morphological processing to complete image segmentation. The invention greatly improves the general adaptation of the algorithm by realizing the self-adaptive adjustment of the key parameters in each link of the segmentation process. Compared with the prior art, the method fully utilizes the information carried by the color abrasive particle image, has clear and complete segmentation process and stable segmentation result, and greatly improves the accuracy of abrasive particle segmentation work.
Description
Technical Field
The invention relates to the technical field of oil abrasive particle analysis, in particular to a universal ferrographic abrasive particle image segmentation method.
Background
Ferrographic image segmentation is a key link for developing an oil abrasive particle analysis technology, segmentation images are the premise for developing automatic abrasive particle identification work, and the accuracy of the automatic abrasive particle identification work is directly influenced by the segmentation precision and the segmentation efficiency. Under the influence of the complexity of the operation condition of equipment, abrasive particles in oil liquid present shape diversity and texture diversity, and are highlighted in the aspects of particle size, thickness, edges, surfaces, background and the like, so that the abrasive particle images have obvious intra-class difference and inter-class difference. Under the influence of obvious differences of abrasive particle images, when the existing ferrographic abrasive particle image segmentation technology is applied to different abrasive particle image segmentation works, the over-segmentation or under-segmentation phenomenon is always in an uncontrollable state, and processes such as morphological post-processing and the like need to be introduced on the basis of a segmentation algorithm to greatly correct the segmentation result, so that the quality of the segmentation result is greatly dependent on interactive processing of segmentation personnel. Therefore, how to improve the adaptive degree of the segmentation algorithm and reduce the interactive influence of personnel on the segmentation process to the maximum extent is one of the key problems to be solved in the field of oil abrasive particle image segmentation and also one of the key problems in realizing the automation of the abrasive particle identification workflow.
At present, in the automatic identification link of the abrasive particles, methods such as an Otsu threshold value and color clustering are mainly adopted for segmenting the ferrographic abrasive particle image, the utilization rate of the bearing information of the abrasive particle image is low, the segmentation result is unstable, and a set of oil abrasive particle image segmentation method with high information utilization rate, reasonable flow and strong self-adaption degree is lacked to comprehensively improve the automation degree of the abrasive particle identification flow.
Disclosure of Invention
In view of this, the present invention provides an oil abrasive particle image segmentation method, so as to solve the problems of unstable segmentation result, excessive dependence on interactive processing, and the like existing in the ferrographic abrasive particle image segmentation method in the prior art, and the method can complete batch segmentation of abrasive particle images by using the set of processes, thereby improving the automation degree of the abrasive particle identification work.
The invention relates to an oil abrasive particle image segmentation method, which comprises the following steps:
1) carrying out gray morphological reconstruction on an image to be segmented, extracting a gradient image, and correcting the gradient image based on an H-minima technology, wherein H and a correction value H' thereof are set as follows:
H’=β·H(0<β≤1) (2)
in the formula: m is a group of 0 、M 1 、M 2 Respectively representing the mean value of the corrected gradient image, the mean value of the local minimum value and the mean value of the local maximum value, wherein beta is a fixed value correction factor;
carrying out watershed change on the corrected gradient image to complete one-time segmentation of the abrasive particle image and obtain a watershed homogeneous region;
2) completing the color feature extraction of watershed homogeneous regions in a Lab color space: converting the abrasive particle image from an RGB space to a Lab space, respectively compressing the L channel image, the a channel image and the b channel image to N color grades, obtaining color values of all homogeneous regions in the watershed segmentation image, sequentially extracting color distribution histograms of all the regions, and performing normalization processing on the color distribution histograms;
3) finishing the extraction of texture features of watershed homogeneous regions based on an LBP map: determining P sampling points in an area with the radius of R, acquiring an LBP map, calculating a texture distribution histogram of each homogeneous area in a watershed segmentation image based on the LBP map, and carrying out normalization processing on the histogram;
4) respectively completing similarity measurement of color and texture normalized distribution histograms based on Bhattacharyya coefficients, and obtaining a homogeneous region R m And R n Color similarity p between them color Similarity to texture ρ LBP Is calculated as follows:
in the formula:are respectively a region R m The color normalized histogram and the texture normalized histogram of (1);are respectively a region R n Color normalization ofA histogram and a texture normalized histogram;
in the process of feature fusion, a region R is defined m And R n The comprehensive similarity matrix w of (a) is:
w(R m ,R n )=w color (R m ,R n )·ρ color (R m ,R n )+w LBP (R m ,R n )·ρ LBP (R m ,R n ) (7)
the comprehensive similarity matrix W of the watershed homogeneous region is as follows:
5) automatically marking seed regions based on the similarity matrix, comprising the steps of:
a. selection and labeling of background seed regions: defining the homogeneous region with the largest product after watershed segmentation as a background seed region R b The region is marked 1, i.e. L (R) b )=1;
b. Selection and marking of foreground seed regions: tracking the edge contour of a target area, judging the mark values of pixels outside the contour, wherein the mark values corresponding to the pixels outside the contour are all adjacent to the target area, so that the mark values of adjacent areas are obtained, the mark values of homogeneous areas after being segmented by a watershed algorithm are set to form a set s, and the specific judgment process of a foreground seed area is as follows by taking the adjacent areas as reference:
extracting a background region R b Is marked with a value of R v ,And is provided withForeground seed region at R v Selecting; s represents a mark value of a homogeneous region after the watershed algorithm segmentation, and is a set, v is a subset of s, and v does not contain b elements;
② obtaining R in turn v Of adjacent region, denoted as R vr ,Extracting R from the similarity matrix W v Similarity value w (R) with its neighboring region v ,R r );
Thirdly, sequentially calculating the region R according to the similarity w v The node degree matrix D of (2), the corresponding region mark value v when the matrix D takes the maximum value 0 Defining the region as a foreground region seed point, and marking the region as 0;
the calculation formula in the foreground seed region marking process is as follows:
v 0 =max(D(R v )) (9)
6) merging of watershed homogeneous regions based on region merging criterion
Based on background seed region R b And foreground seed regionRegion R i (i belongs to s, i is not equal to b, i is not equal to v 0 ) The merging rule of (1) is:
7) and after the merging work is finished, performing unified morphological post-processing on the merged image to finish the image segmentation work.
The invention has the beneficial effects that:
compared with the prior art, the oil abrasive particle image segmentation method has the following advantages:
1) the method makes full use of the information carried by the color image of the wear particles, so that the single abrasive particle image can obtain the best compromise between the segmentation efficiency and the segmentation precision, and the stability of the segmentation effect is ensured.
2) The method provides a set of complete abrasive particle image segmentation process, optimizes each segmentation link to different degrees, fully utilizes the advantages of a morphological algorithm, furthest reduces the dependence degree on a morphological adjustment effect, realizes the self-adaptive adjustment of the segmentation algorithm by introducing a self-adaptive H value in the H-minima technology, introducing self-adaptive weight in the characteristic fusion process, realizing the automatic selection and marking of seed regions and other methods, avoids the interactive processing in the segmentation process, and provides possibility for batch segmentation of abrasive particle images.
3) The method is suitable for various abrasive particles generated under a typical abrasion mechanism, and meanwhile, as the extraction methods of the color characteristics and the texture characteristics have scale invariance, the method is also suitable for abrasive particle images under various scales and has good universality.
Drawings
FIG. 1 is a flow chart of oil abrasive particle image segmentation;
FIG. 2 is an image of abrasive particles to be segmented;
FIG. 3 is a watershed marked image extracted after morphological reconstruction of a gray level image to be segmented;
FIG. 4 is a watershed labeled image extracted after a gradient image is corrected based on the H-minima technology;
FIG. 5 is a color normalized histogram of watershed homogeneous regions 1 and 39 of a grain image to be segmented extracted based on Lab color space;
FIG. 6 is a texture normalized histogram of watershed homogeneous regions 1 and 39 of an image of a grit to be segmented extracted based on LBP maps;
FIG. 7 is an image after watershed homogenous region merging based on region merging criteria;
FIG. 8 is an image after correction of the merged image based on morphological post-processing;
FIG. 9 is a flow chart of a unified morphological post-processing of the merged images.
Detailed Description
The invention is further described below with reference to the figures and examples.
As shown in the figure, the oil abrasive particle image segmentation method of the embodiment includes the following steps:
1) carrying out gray morphological reconstruction on an image to be segmented, extracting a gradient image, and correcting the gradient image based on an H-minima technology, wherein H and a correction value H' thereof are set as follows:
H’=β·H(0<β≤1) (2)
in the formula: m 0 、M 1 、M 2 Respectively representing the mean value of the corrected gradient image, the mean value of the local minimum value and the mean value of the local maximum value, wherein beta is a fixed value correction factor, and is taken to be 0.5 on the basis of statistical analysis of the test sample;
and carrying out watershed change on the corrected gradient image, completing one-time segmentation of the abrasive particle image, and obtaining a watershed homogeneous region.
2) Completing the color feature extraction of watershed homogeneous regions in a Lab color space: converting the abrasive grain image from an RGB space to a Lab space, compressing the L channel image, the a channel image and the b channel image to N color levels, where N is 16 in this embodiment, obtaining color values of each homogeneous region in the watershed segmented image, sequentially extracting color distribution histograms of each region, and performing normalization processing on the color distribution histograms.
3) Finishing the extraction of texture features of watershed homogeneous regions based on an LBP map: determining P sampling points in an area with the radius of R, wherein R is 2, and P is 8 in the embodiment; and acquiring an LBP map, calculating a texture distribution histogram of each homogeneous region in the watershed segmentation image based on the LBP map, and carrying out normalization processing on the histogram.
4) And respectively completing similarity measurement of the normalized distribution histograms of the color and the texture based on the Bhattacharyya coefficient, and realizing proportional distribution of the weight of the characteristic index according to the similarity of the characteristic in the characteristic fusion process. Homogeneous region R m And R n Color similarity p between them color Similarity to texture ρ LBP Is calculated as follows:
in the formula:are respectively a region R m The color normalized histogram and the texture normalized histogram of (1);are respectively a region R n The color normalized histogram and the texture normalized histogram of (1).
In the process of feature fusion, a region R is defined m And R n The comprehensive similarity matrix w of (a) is:
w(R m ,R n )=w color (R m ,R n )·ρ color (R m ,R n )+w LBP (R m ,R n )·ρ LBP (R m ,R n ) (7)
the comprehensive similarity matrix W of the watershed homogeneous region is as follows:
5) automatically marking seed regions based on the similarity matrix, comprising the steps of:
a. selection and labeling of background seed regions: defining the homogeneous region with the largest product after watershed segmentation as a background seed region R b The region is marked 1, L (R) b )=1;
b. Selection and marking of foreground seed regions: tracking the edge contour of a target area, judging the marking values of pixels outside the contour, wherein the marking values corresponding to the pixels outside the contour are all adjacent to the target area, so that the marking values of adjacent areas are obtained, the marking values of homogeneous areas after being segmented by a watershed algorithm are set to form a set s, and the specific judgment process of a foreground seed area is as follows by taking adjacent areas as reference:
extracting a background region R b Is marked with a value of R v ,And isForeground seed region at R v Selecting; s represents a mark value of a homogeneous region after the watershed algorithm segmentation, and is a set, v is a subset of s, and v does not contain b elements;
(ii) sequentially obtaining R v Is marked as R vr ,Extracting R from the similarity matrix W v Similarity value w (R) with its neighboring region v ,R r );
Thirdly, sequentially calculating the region R according to the similarity w v The node degree matrix D of (2), the corresponding region mark value v when the matrix D takes the maximum value 0 Defining the region as a foreground region seed point, and marking the region as 0;
the calculation formula in the foreground seed region marking process is as follows:
v 0 =max(D(R v )) (9)
6) merging of watershed homogeneous regions based on region merging criterion
Based on background seed region R b And foreground seed regionRegion R i (i belongs to s, i is not equal to b, i is not equal to v 0 ) The merging rule of (1) is:
7) and after the merging work is finished, performing unified morphological post-processing on the merged image to finish the image segmentation work. The flow of the unified morphological post-processing of the merged images is as follows:
firstly, processing the merged binary image based on morphological erosion operation for eliminating false segmentation lines appearing in the background, wherein in the step, a circular structural element is selected, and the size se is equal to 1;
filling holes in the binary image processed in the first step for correcting mistakenly segmented pixel points on the surface of the abrasive particles due to severe gray change;
thirdly, deleting a small area of the binary image processed in the second step to ensure that only the segmented view field contains target abrasive particles;
and fourthly, performing morphological opening operation on the binary image processed in the third step to eliminate segmentation noise points and smooth the segmentation contour, thereby completing the segmentation of the abrasive particle image. In this embodiment, a circular structural element is selected at this step, and the size se is 3.
Finally, the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all of them should be covered in the claims of the present invention.
Claims (1)
1. An oil abrasive particle image segmentation method is characterized by comprising the following steps: the method comprises the following steps:
1) carrying out gray morphological reconstruction on an image to be segmented, extracting a gradient image, and correcting the gradient image based on an H-minima technology, wherein H and a correction value H' thereof are set as follows:
H’=β·H,0<β≤1 (2)
in the formula: m 0 、M 1 、M 2 Respectively representing the mean value of the corrected gradient image, the mean value of the local minimum value and the mean value of the local maximum value, wherein beta is a fixed value correction factor;
carrying out watershed change on the corrected gradient image to complete one-time segmentation of the abrasive particle image and obtain a watershed homogeneous region;
2) completing the extraction of the color characteristics of watershed homogeneous regions in a Lab color space: converting the abrasive particle image from an RGB space to a Lab space, respectively compressing the L channel image, the a channel image and the b channel image to N color grades, obtaining color values of all homogeneous regions in the watershed segmentation image, sequentially extracting color distribution histograms of all the regions, and performing normalization processing on the color distribution histograms;
3) finishing the extraction of texture features of watershed homogeneous regions based on an LBP map: determining P sampling points in an area with the radius of R, acquiring an LBP map, calculating a texture distribution histogram of each homogeneous area in a watershed segmentation image based on the LBP map, and carrying out normalization processing on the histogram;
4) based on Bhattacharyya coefficient, similarity measurement of color and texture normalized distribution histograms is respectively completed, and homogeneous regions R m And R n Color similarity ρ between color Similarity to texture ρ LBP Is calculated as follows:
in the formula:are respectively a region R m The color normalized histogram and the texture normalized histogram of (1);are respectively a region R n The color normalized histogram and the texture normalized histogram of (1);
in the process of feature fusion, a region R is defined m And R n The comprehensive similarity matrix w of (a) is:
w(R m ,R n )=w color (R m ,R n )·ρ color (R m ,R n )+w LBP (R m ,R n )·ρ LBP (R m ,R n ) (7)
the comprehensive similarity matrix W of the watershed homogeneous region is as follows:
5) automatically marking seed regions based on the similarity matrix, comprising the steps of:
a. selection and labeling of background seed regions: defining the homogeneous region with the largest product after watershed segmentation as a background seed region R b The region is marked 1, L (R) b )=1;
b. Selection and marking of foreground seed regions: tracking the edge contour of a target area, judging the marking values of pixels outside the contour, wherein the marking values corresponding to the pixels outside the contour are all adjacent to the target area, so that the marking values of adjacent areas are obtained, the marking values of homogeneous areas after being segmented by a watershed algorithm are set to form a set s, and the specific judgment process of a foreground seed area is as follows by taking adjacent areas as reference:
firstly, extracting a background region R b Is marked with a value of R v ,And isForeground seed region at R v Selecting; s represents a mark value of a homogeneous region after the watershed algorithm segmentation, and is a set, v is a subset of s, and v does not contain b elements;
② obtaining R in turn v Is marked as R vr ,Extracting R from the similarity matrix W v Similarity value w (R) with its adjacent region v ,R r );
Thirdly, sequentially calculating the region R according to the similarity w v The node degree matrix D of (1), the corresponding region mark value v when the matrix D takes the maximum value 0 Defining the region as a foreground region seed point, and marking the region as 0;
the calculation formula in the foreground seed region marking process is as follows:
v 0 =max(D(R v )) (9)
6) merging of watershed homogeneous regions based on region merging criterion
Based on background seed region R b And foreground seed regionRegion R i The merging rule of (1) is:
wherein R is i The subscript i belongs to s, i is not equal to b, i is not equal to v 0 ;
7) And after the merging work is finished, performing unified morphological post-processing on the merged image to finish the image segmentation work.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811559431.6A CN109584253B (en) | 2018-12-20 | 2018-12-20 | Oil abrasive particle image segmentation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811559431.6A CN109584253B (en) | 2018-12-20 | 2018-12-20 | Oil abrasive particle image segmentation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109584253A CN109584253A (en) | 2019-04-05 |
CN109584253B true CN109584253B (en) | 2022-08-30 |
Family
ID=65930126
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811559431.6A Active CN109584253B (en) | 2018-12-20 | 2018-12-20 | Oil abrasive particle image segmentation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109584253B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110223299B (en) * | 2019-06-12 | 2021-06-18 | 重庆邮电大学 | Abrasive particle segmentation method based on deposition process |
CN111008949B (en) * | 2019-08-16 | 2021-09-14 | 苏州喆安医疗科技有限公司 | Soft and hard tissue detection method for tooth image |
CN111445485B (en) * | 2020-03-10 | 2023-05-23 | 西安工业大学 | Online abrasive particle image data processing method |
CN111681244B (en) * | 2020-05-29 | 2022-06-21 | 山东大学 | Blade image segmentation method, system, equipment and storage medium |
CN112183556B (en) * | 2020-09-27 | 2022-08-30 | 长光卫星技术股份有限公司 | Port ore heap contour extraction method based on spatial clustering and watershed transformation |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105809673A (en) * | 2016-03-03 | 2016-07-27 | 上海大学 | SURF (Speeded-Up Robust Features) algorithm and maximal similarity region merging based video foreground segmentation method |
CN106599793A (en) * | 2016-11-21 | 2017-04-26 | 江苏大学 | Marked watershed segmentation-based steel grain boundary automatic extraction method |
CN107481241A (en) * | 2017-08-24 | 2017-12-15 | 太仓安顺财务服务有限公司 | A kind of color image segmentation method based on mixed method |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7327880B2 (en) * | 2004-03-12 | 2008-02-05 | Siemens Medical Solutions Usa, Inc. | Local watershed operators for image segmentation |
WO2016001223A1 (en) * | 2014-06-30 | 2016-01-07 | Ventana Medical Systems, Inc. | Detecting edges of a nucleus using image analysis |
-
2018
- 2018-12-20 CN CN201811559431.6A patent/CN109584253B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105809673A (en) * | 2016-03-03 | 2016-07-27 | 上海大学 | SURF (Speeded-Up Robust Features) algorithm and maximal similarity region merging based video foreground segmentation method |
CN106599793A (en) * | 2016-11-21 | 2017-04-26 | 江苏大学 | Marked watershed segmentation-based steel grain boundary automatic extraction method |
CN107481241A (en) * | 2017-08-24 | 2017-12-15 | 太仓安顺财务服务有限公司 | A kind of color image segmentation method based on mixed method |
Non-Patent Citations (2)
Title |
---|
"Watershed Image Segmentation Algorithm Base on Particle Swarm and Region Growing";SUN Hui-jie;《2015 International Conference on Intelligent Computing and Internet of Things (IClT)》;20150525;第1-4页 * |
"基于改进分水岭和区域合并的彩色图像分割";李占波 等;《计算机工程与设计》;20140731;第1-5页 * |
Also Published As
Publication number | Publication date |
---|---|
CN109584253A (en) | 2019-04-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109584253B (en) | Oil abrasive particle image segmentation method | |
CN109377485B (en) | Machine vision detection method for instant noodle packaging defects | |
CN112365494B (en) | Ore material image segmentation method based on deep learning prediction edge | |
CN114862862A (en) | Pump body cold shut defect identification method and system based on image processing | |
CN115331119B (en) | Solid waste identification method | |
CN106404793A (en) | Method for detecting defects of bearing sealing element based on vision | |
CN111652213A (en) | Ship water gauge reading identification method based on deep learning | |
CN109598681B (en) | No-reference quality evaluation method for image after repairing of symmetrical Thangka | |
US11037299B2 (en) | Region merging image segmentation algorithm based on boundary extraction | |
CN112734729B (en) | Water gauge water level line image detection method and device suitable for night light supplement condition and storage medium | |
CN116993731B (en) | Shield tunneling machine tool bit defect detection method based on image | |
CN116630971B (en) | Wheat scab spore segmentation method based on CRF_Resunate++ network | |
CN116524196B (en) | Intelligent power transmission line detection system based on image recognition technology | |
CN109341524A (en) | A kind of optical fiber geometric parameter detection method based on machine vision | |
CN111738931B (en) | Shadow removal algorithm for aerial image of photovoltaic array unmanned aerial vehicle | |
CN111210447B (en) | Hematoxylin-eosin staining pathological image hierarchical segmentation method and terminal | |
Mubashshira et al. | An unsupervised approach for road surface crack detection | |
CN107729863B (en) | Human finger vein recognition method | |
CN113435285A (en) | Automatic analysis method and system for chromosome karyotype of hematological tumor | |
CN115619799B (en) | Grain image segmentation method and system based on transfer learning | |
CN112215303A (en) | Image understanding method and system based on self-learning attribute | |
CN113763404B (en) | Foam image segmentation method based on optimization mark and edge constraint watershed algorithm | |
CN115187788A (en) | Crop seed automatic counting method based on machine vision | |
CN113850792A (en) | Cell classification counting method and system based on computer vision | |
Abrol et al. | An automated segmentation of leukocytes using modified watershed algorithm on peripheral blood smear images |
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 |