CN105160675A - Automatic segmentation method for powdery mildew spore image - Google Patents
Automatic segmentation method for powdery mildew spore image Download PDFInfo
- Publication number
- CN105160675A CN105160675A CN201510544493.XA CN201510544493A CN105160675A CN 105160675 A CN105160675 A CN 105160675A CN 201510544493 A CN201510544493 A CN 201510544493A CN 105160675 A CN105160675 A CN 105160675A
- Authority
- CN
- China
- Prior art keywords
- image
- spore
- pixel
- segmentation
- region
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
Abstract
Disclosed is an automatic segmentation method for a powdery mildew spore image. The method comprises: designing and processing an adhered spore binary image in combination with morphological basic operation; defining a calculation matrix, and generating a distance transformation matrix and a global variable for counting with each erosion process; performing erosion operation on the binary image by using morphological erosion operation, scanning the whole image, and marking disappeared pixel points; circulating until all communicated regions are eroded and disappeared; when processing all the marked regions, obtaining all seed points; and calling a watershed algorithm for an obtained distance image, immersing spores from big to small according to pixel gray values in the communicated regions, obtaining a segmentation line, performing spore segmentation, and disconnecting adhered edge parts, thereby improving the effect of finally performing watershed segmentation on the adhered spore image. According to the method, the process for calculating the seed points is more accurate to judge, the jump of the gray scale gradient between the pixel points of the obtained distance image is relatively stable, and cell segmentation is also more accurate.
Description
Technical field
The present invention can obtain adhesion Powdery Mildew spore image and compare the more accurate segmentation effect of commonsense method, belongs to digital image processing techniques field.
Background technology
Powdery mildew is one of Major Diseases of plant disease, its main harm plant leaf blade, stem, petiole, bud and petal, and the plant that is then injured there will be chlorisis, withered and yellow, shrinkage, spire distortion, causes the bad developmental deformity of plant strain growth, can not yield positive results.Wheat powdery mildew is one of Major Diseases of wheat, except some torrid areas, all has distribution all over the world.Before 1900, the states such as the Britain in Europe, France, Ireland, Belgium, Denmark, Holland, Switzerland and Finland appear in the newspapers.Though the U.S., Canada, Mexico have and to occur but smaller.In recent years, wheat powdery mildew has become one of Major Diseases of China's wheat, that onset area or occurring degree all maintain a higher level, average annual onset area is more than 6,000,000 hectares, general morbidity can cause about 10% underproduction, time serious, the underproduction can reach more than 50%, and even No kernels or seeds are gathered, as in a year of scarcity for some high sense kind.So wheat powdery mildew has become one of important normal venereal disease evil in China's Wheat Production, also become one of the main restricting factor of China's improving yield of wheat, stable yields, high-quality at present.
According to statistics, in the 742 kinds of corps diseases in the whole nation, about have 60% to infect crop by germ spore through air-flow propagation and form harm.The gas borne disease that wheat powdery mildew is caused by obligatory parasitism fungi-wheat powdery mildew spore exactly.Wheat powdery mildew belongs to sac fungus, and can only grow in the host tissue of living, and have very strict specialization to host, generally only infect wheat, so wheat germ can not infect barley, large wheat powdery mildew does not also infect wheat.Wheat powdery mildew spreads in host surfaces with mycelium, stretches in epidermal cells of host draw nutrition by means of only haustorium.On mycelia raw conidiophore and conidium, conidium is oval, length be the 1.25-1.75 of width doubly, minority reaches nearly two times, unit cell is colourless, and conidiophore is upright, vertically bears from mycelium, not branch, colourless, top produces the conidium of bunchiness, number 10 ~ 20, from top, maturation comes off gradually downwards, propagates to cause infect and plant disease epidemic again with air-flow.Conidium all can sprout within the scope of 0.5 ~ 30 DEG C, and the optimum temperature of sprouting is 10 ~ 18 DEG C.Very wide to the accommodation of humidity, under optimum temperature condition, relative humidity 0 ~ 100% all can be sprouted, but humidity is larger, and germination rate is higher.Therefore under wheat powdery mildew interacts for a long time with host in the geographical ecologic environment of difference, can form different biological strains, Toxicity Variation is very fast, and when environmental baseline is suitable, wheat powdery mildew will spread in a large number, causes the generation of wheat powdery mildew.
Current powdery mildew preventing control method carries out the observation of white powder disease mainly through artificial visually examine, field investigation etc., observe underaging, labour intensity is large, digitized degree is low.Also need timing regular sampling germ spore to be taken back laboratory at biology microscope Microscopic observation, manually record is carried out to spore quantity.This ways and means backwardness relatively, the present invention is directed to the important step of image procossing--and the range conversion step of Iamge Segmentation link is improved, and improves segmentation effect, the first step improves prevention and control.
Summary of the invention
Fundamental purpose of the present invention is the degree of accuracy improving the range image adopted in Iamge Segmentation pre-treatment step, namely obtains the slope that gradation of image is good.From image target edge to the center of target, grey scale change is stable, obvious.Significantly adhesion part and each spore center can be distinguished and come.Effective raising is for the final segmentation effect of adhesion Powdery Mildew spore.
The automatic division method of Powdery Mildew spore image, this method have employed morphological erosion computing and combines and carry out counting to each pixel and come.This method uses MATLAB write and realize, and devises the distance transformation method based on morphology operations.This Iamge Segmentation pre-treatment step of conversion of adjusting the distance is improved, and improves the final segmentation effect to adhesion Powdery Mildew spore.In order to obtain range image, first full figure is scanned, with each target area to be split of mode record of filling or mark calculates (spore namely before segmentation), then 3 × 3 structural elements are used to carry out the process of morphological erosion to each target area, (passing marker region, passing marker region again after corrosion, instead of full figure), following three kinds of situations may be there are through scanning after marked region is corroded:
(1) this region is still a connected region.
(2) this region is divided into multiple connected region.
(3) this zones vanishes.
For (1), corrosion will be proceeded, until there is the situation of (2) or (3).
For (2), adopted by marked region the chained list in data structure to preserve, after emerging N number of marked region being inserted into respectively the chained list current node in data structure, sailing again and ruin current node.
For (3), the pixel in this region, namely as Seed Points, continues the next marked region of process after recording Seed Points.
After the complete all marked regions of such operation, the Seed Points of watershed algorithm is just all found complete, and process spore figure being become range conversion figure also completes in process above, namely only need to record the pixel be at every turn corroded, and give these pixels a value, this value is the number of times that current connected region is corroded.
The concrete technical scheme that the present invention takes is as follows:
S1 is for the pre-treatment step of existing watershed segmentation methods, i.e. range conversion link, in order to obtain better gray ramp than general distance transformation method (as Euclidean distance, chessboard distance, city distance etc.), obtain Seed Points and range image accurately, combining form fundamental operation of the present invention, designs the method processed adhesion spore bianry image.Definition compute matrix, generate range conversion matrix with along with corroding the global variable carrying out counting at every turn.Use morphological erosion computing, carry out etching operation to bianry image, scanning full figure, marks the pixel disappeared.Circulation aforesaid operations, is labeled as N, until all connected regions are etched to disappearance to the pixel that the N time disappears.When processing all marked regions, also obtain all Seed Points.Watershed algorithm is called to the range image obtained, according to the submergence spore from big to small of the grey scale pixel value in connected region, obtain cut-off rule, then carry out spore segmentation, disconnect in the marginal portion of adhesion, improve final effect of adhesion spore image being carried out to watershed segmentation.
S2 this method uses gray_image=rgb2gray (rgb_image) instruction that the coloured image of input is converted to gray level image, then 1-im2bw (filter_image is re-used after using level=graythresh (gray_image) instruction to become bianry image according to adaptively selected suitable threshold value, level) method is negated, and finally obtains required bianry image.Wherein greythresh (gray_image) instruction uses the appropriate threshold of maximum variance between clusters Automatic-searching gray level image, finds image suitable threshold value.All different threshold values can be obtained for often opening image.Maximum variance between clusters is the gamma characteristic according to image, image is divided into background and target two parts.Inter-class variance between background and target is larger, illustrates that the two-part difference of composing images is less, when partial target mistake is divided into background or part background mistake to be divided into target that two parts difference all can be caused to diminish.Last in order to subsequent operation, use uint8 instruction map image type.
S3 uses morphological erosion computing, carries out etching operation to bianry image, and scanning full figure, marks the pixel disappeared.The present invention uses imerode instruction in Matlab to carry out etching operation to image, uses 3 × 3 unit matrixs as structural element matrix, namely
Middle 1 is structural element center.Use 3 × 3 unit matrixs to be each pixel with structural element scan image as structural element matrix etching operation, make and operation of the bianry image of structural element and its covering: if be all 1, this pixel point value of result images is 1, otherwise is 0.
S4 circulation aforesaid operations, is labeled as N, until connected regions all in bianry image is etched to disappearance to the pixel that the N time disappears.
When processing all marked regions, also obtain all Seed Points.
Bianry image is transformed to the gray level image that from each connected region edge to Seed Points process, numerical value increases gradually, and the gray-scale value of background is 0.Gray level image graded is stablized, and it is inner that Seed Points is accurately positioned at each connected region.The erosion operation that morphology uses is as follows:
Set A is written as A Θ B by set B corrosion, is defined as
S5 defines compute matrix, generate range conversion matrix with along with corroding the global variable carrying out counting at every turn.After compute matrix corrodes at every turn, in the range conversion matrix generated, counting global variable currency is used to mark accordingly to the pixel becoming 0 from 1.
S6 scans full figure and is and sues for peace to full images vegetarian refreshments, if result is not 0, then continues cycling, if be 0 i.e. terminating operation, obtains final range image.The present invention uses the q=sum (sum (Bw)) in Matlab to obtain the cumulative sum of a pixel in compute matrix.
S7 adopts dividing ridge method to the range image obtained, and according to the submergence spore from big to small of the grey scale pixel value in connected region, obtains cut-off rule, then carries out spore segmentation.Final realization, to the segmentation of adhesion spore, makes the marginal portion of adhesion disconnect.The present invention does not grow adhesion edge and does not have the adhesion spore of lap respond well.
Advantage of the present invention:
(1) degree of accuracy that common method asks for Seed Points is not high, and segmentation effect is not ideal enough.It is more accurate that this method asks Seed Points process to judge, between the range image pixel obtained, shade of gray jumps more stable, and cell segmentation is also more accurate.
(2) the present invention design distance transformation method effective, realize easy.
Accompanying drawing explanation
Fig. 1 is bianry image connected region matrix.
Fig. 2 is the two values matrix after etching operation.
Fig. 3 is the former figure of Powdery Mildew spore image (in figure, arrow indication is Powdery Mildew spore).
Fig. 4 is Powdery Mildew spore image former figure segmentation effect figure.
Embodiment
For three kinds of different bianry image connected regions, the outermost layer pixel point set disappeared is marked, and determine Seed Points according to different situations after each corrosion.
The segmentation effect final based on morphologic distance transformation method of the present invention's design preferably resolves the problem of adhesion between target.
Claims (1)
1. the automatic division method of Powdery Mildew spore image, this method have employed morphological erosion computing and combines and carry out counting to each pixel and come; This method uses MATLAB write and realize, and devises the distance transformation method based on morphology operations; This Iamge Segmentation pre-treatment step of conversion of adjusting the distance is improved, and improves the final segmentation effect to adhesion Powdery Mildew spore;
In order to obtain range image, first full figure is scanned, spore before namely splitting with each target area to be split of mode record of filling or mark calculates, then 3 × 3 structural elements are used to carry out the process of morphological erosion to each target area, passing marker region again after corrosion,, after marked region is corroded, following three kinds of situations may be there are through scanning in passing marker region, instead of full figure:
(1) this region is still a connected region;
(2) this region is divided into multiple connected region;
(3) this zones vanishes;
For (1), corrosion will be proceeded, until there is the situation of (2) or (3);
For (2), adopted by marked region the chained list in data structure to preserve, after emerging N number of marked region being inserted into respectively the chained list current node in data structure, sailing again and ruin current node;
For (3), the pixel in this region, namely as Seed Points, continues the next marked region of process after recording Seed Points;
After the complete all marked regions of such operation, the Seed Points of watershed algorithm is just all found complete, and process spore figure being become range conversion figure also completes in process above, namely only need to record the pixel be at every turn corroded, and give these pixels a value, this value is the number of times that current connected region is corroded;
It is characterized in that: the implementation procedure of the method is as follows,
S1 is for the pre-treatment step of existing watershed segmentation methods, i.e. range conversion link, in order to obtain better gray ramp than general distance transformation method, obtain Seed Points and range image accurately, the fundamental operation of this method combining form, designs the method processed adhesion spore bianry image; Definition compute matrix, generate range conversion matrix with along with corroding the global variable carrying out counting at every turn; Use morphological erosion computing, carry out etching operation to bianry image, scanning full figure, marks the pixel disappeared; Circulation aforesaid operations, is labeled as N, until all connected regions are etched to disappearance to the pixel that the N time disappears; When processing all marked regions, also obtain all Seed Points; Watershed algorithm is called to the range image obtained, according to the submergence spore from big to small of the grey scale pixel value in connected region, obtain cut-off rule, then carry out spore segmentation, disconnect in the marginal portion of adhesion, improve final effect of adhesion spore image being carried out to watershed segmentation;
S2 this method uses gray_image=rgb2gray (rgb_image) instruction that the coloured image of input is converted to gray level image, then 1-im2bw (filter_image is re-used after using level=graythresh (gray_image) instruction to become bianry image according to adaptively selected suitable threshold value, level) method is negated, and finally obtains required bianry image; Wherein greythresh (gray_image) instruction uses the appropriate threshold of maximum variance between clusters Automatic-searching gray level image, finds image suitable threshold value; All different threshold values can be obtained for often opening image; Maximum variance between clusters is the gamma characteristic according to image, image is divided into background and target two parts; Inter-class variance between background and target is larger, illustrates that the two-part difference of composing images is less, when partial target mistake is divided into background or part background mistake to be divided into target that two parts difference all can be caused to diminish; Last in order to subsequent operation, use uint8 instruction map image type;
S3 uses morphological erosion computing, carries out etching operation to bianry image, and scanning full figure, marks the pixel disappeared; This method uses imerode instruction in Matlab to carry out etching operation to image, uses 3 × 3 unit matrixs as structural element matrix, namely
Middle 1 is structural element center; Use 3 × 3 unit matrixs to be each pixel with structural element scan image as structural element matrix etching operation, make and operation of the bianry image of structural element and its covering: if be all 1, this pixel point value of result images is 1, otherwise is 0;
S4 circulation aforesaid operations, is labeled as N, until connected regions all in bianry image is etched to disappearance to the pixel that the N time disappears;
When processing all marked regions, also obtain all Seed Points;
Bianry image is transformed to the gray level image that from each connected region edge to Seed Points process, numerical value increases gradually, and the gray-scale value of background is 0; Gray level image graded is stablized, and it is inner that Seed Points is accurately positioned at each connected region; The erosion operation that morphology uses is as follows:
Set A is written as A Θ B by set B corrosion, is defined as
S5 defines compute matrix, generate range conversion matrix with along with corroding the global variable carrying out counting at every turn; After compute matrix corrodes at every turn, in the range conversion matrix generated, counting global variable currency is used to mark accordingly to the pixel becoming 0 from 1;
S6 scans full figure and is and sues for peace to full images vegetarian refreshments, if result is not 0, then continues cycling, if be 0 i.e. terminating operation, obtains final range image; This method uses the q=sum (sum (Bw)) in Matlab to obtain the cumulative sum of a pixel in compute matrix;
S7 adopts dividing ridge method to the range image obtained, and according to the submergence spore from big to small of the grey scale pixel value in connected region, obtains cut-off rule, then carries out spore segmentation; Final realization, to the segmentation of adhesion spore, makes the marginal portion of adhesion disconnect; This method is not grown adhesion edge and is not had the adhesion spore of lap respond well.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510544493.XA CN105160675A (en) | 2015-08-31 | 2015-08-31 | Automatic segmentation method for powdery mildew spore image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510544493.XA CN105160675A (en) | 2015-08-31 | 2015-08-31 | Automatic segmentation method for powdery mildew spore image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105160675A true CN105160675A (en) | 2015-12-16 |
Family
ID=54801517
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510544493.XA Pending CN105160675A (en) | 2015-08-31 | 2015-08-31 | Automatic segmentation method for powdery mildew spore image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105160675A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833200A (en) * | 2017-09-19 | 2018-03-23 | 浙江农林大学 | It is a kind of to detect the independent method and system with adhesion cardiac muscle cell's core region |
| CN107909138A (en) * | 2017-11-14 | 2018-04-13 | 江苏大学 | A kind of class rounded grain thing method of counting based on Android platform |
| CN108052886A (en) * | 2017-12-05 | 2018-05-18 | 西北农林科技大学 | A kind of puccinia striiformis uredospore programming count method of counting |
| CN108918481A (en) * | 2018-04-26 | 2018-11-30 | 南昌大学 | A kind of huve cell system of fluorescence analysis |
| CN112074841A (en) * | 2018-03-30 | 2020-12-11 | 珀金埃尔默健康科学有限公司 | System and method for automatically detecting and segmenting vertebral bodies in 3D images |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091179A (en) * | 2014-07-01 | 2014-10-08 | 北京工业大学 | Intelligent blumeria graminis spore picture identification method |
| CN104484877A (en) * | 2014-12-12 | 2015-04-01 | 山东大学 | AML cell segmentation method based on Meanshift cluster and morphological operations |
-
2015
- 2015-08-31 CN CN201510544493.XA patent/CN105160675A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091179A (en) * | 2014-07-01 | 2014-10-08 | 北京工业大学 | Intelligent blumeria graminis spore picture identification method |
| CN104484877A (en) * | 2014-12-12 | 2015-04-01 | 山东大学 | AML cell segmentation method based on Meanshift cluster and morphological operations |
Non-Patent Citations (2)
| Title |
|---|
| 张艳诚等: "作物病害图像中重叠病斑分离算法", 《农业机械学报》 * |
| 赵小川等: "《MATLAB图像处理-程序实现与模块化仿真》", 31 January 2014, 北京航空航天大学出版社 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833200A (en) * | 2017-09-19 | 2018-03-23 | 浙江农林大学 | It is a kind of to detect the independent method and system with adhesion cardiac muscle cell's core region |
| CN107833200B (en) * | 2017-09-19 | 2021-03-26 | 浙江农林大学 | Method and system for detecting independent and adhesive myocardial cell nucleus area |
| CN107909138A (en) * | 2017-11-14 | 2018-04-13 | 江苏大学 | A kind of class rounded grain thing method of counting based on Android platform |
| CN108052886A (en) * | 2017-12-05 | 2018-05-18 | 西北农林科技大学 | A kind of puccinia striiformis uredospore programming count method of counting |
| CN112074841A (en) * | 2018-03-30 | 2020-12-11 | 珀金埃尔默健康科学有限公司 | System and method for automatically detecting and segmenting vertebral bodies in 3D images |
| CN108918481A (en) * | 2018-04-26 | 2018-11-30 | 南昌大学 | A kind of huve cell system of fluorescence analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105160675A (en) | Automatic segmentation method for powdery mildew spore image | |
| Maddonni et al. | Intra-specific competition in maize: early establishment of hierarchies among plants affects final kernel set | |
| CN104091179B (en) | Intelligent blumeria graminis spore picture identification method | |
| Chung et al. | Spatial genetic structure of clonal and sexual reproduction in populations of Adenophora grandiflora (Campanulaceae) | |
| Bouttier et al. | Vertical root separation and light interception in a temperate tree-based intercropping system of Eastern Canada | |
| CN102542560B (en) | Method for automatically detecting density of rice after transplantation | |
| CN104805023A (en) | Halimasch fungicide preparing method and halimasch fungicide prepared by preparing method | |
| CN111798327A (en) | Construction method and application of wheat yield calculation model based on hyperspectral image | |
| Shan et al. | Assessment of the genetic connectivity between farmed and wild populations of Undaria pinnatifida (Phaeophyceae) in a representative traditional farming region of China by using newly developed microsatellite markers | |
| CN117349764A (en) | Intelligent analysis method for stem cell induction data | |
| Zhai et al. | Impact of recent breeding history on the competitiveness of Chinese maize hybrids | |
| Neustupa et al. | Temperature-related phenotypic plasticity in the green microalga Micrasterias rotata | |
| Emms et al. | The effect of inflorescence size on male fitness: experimental tests in the andromonoecious lily, Zigadenus paniculatus | |
| Weiland et al. | Prediction of long-term canker disease damage from the responses of juvenile poplar clones to inoculation with Septoria musiva | |
| CN116564407B (en) | Method for efficiently predicting chrysanthemum flowering phase based on whole genome selection | |
| Fan et al. | Genotypic variation for potassium efficiency in wild and domesticated watermelons under ample and limited potassium supply | |
| Segarra-Moragues et al. | Rediscovery of Riella alatospora (Riellaceae, Sphaerocarpales), an aquatic, South African endemic liverwort previously known from a now largely transformed type locality | |
| Renner et al. | Three modes of heterochrony explain lobule diversity in Radula subgenus Cladoradula (Porellales: Jungermanniopsida), a small lineage of early land plants today | |
| CN111919745B (en) | Soybean variety optimization auxiliary screening method based on soybean appearance character linkage inheritance | |
| Shi et al. | Optimizing planting density and irrigation depth of hybrid maize seed production under limited water availability | |
| CN113170669B (en) | Pre-identification and selection method of wheat material with high water-saving index | |
| CN114549881A (en) | Wheat early stem tiller number estimation method based on regional gradual change vegetation index | |
| Christopher | Disease identification in spinach leaves | |
| Sultana et al. | Identification of Ciboria carunculoides RS103V, a fungus causing popcorn disease on mulberry fruits in Korea | |
| CN105261045B (en) | Digital method for rapidly evaluating loss degree of alfalfa diseases and insect pests |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151216 |