CN103593840B - Method for detecting phenotype of Arabidopsis - Google Patents

Abstract

The invention relates to a method for detecting phenotype of Arabidopsis. The method specifically comprises the steps of placing a calibration board in a planting pot of the Arabidopsis, and collecting RGB images of the Arabidopsis by using a camera; carrying out pre-processing on the collected images so as to realize automatic correction and calibration for the images, wherein the image correction is carried out for correcting distortion of the images, and the image calibration is carried out for acquiring the real size of a unit pixel; carrying out segmentation on the pre-processed images, segmenting the Arabidopsis from the background, and extracting the Arabidopsis from the images; extracting phenotype parameters such as the total planting area, the number of leaf blades of the Arabidopsis and the symmetry of the Arabidopsis after the Arabidopsis images are segmented. Arabidopsis phenotype parameters extracted through the method provided by the invention not only can quantitatively describe growing conditions of the Arabidopsis, but also can be applied to the study of gene functions of the Arabidopsis, that is, differences of Arabidopsis with different genes in size and regional form are described through the phenotype parameters, thereby being capable of inferring functions of different genes and influences imposed on Arabidopsis plants.

Description

A kind of arabidopsis Phenotypic examination method

Technical field

The present invention relates to the detection method of a kind of arabidopsis phenotype, utilize the picture of collected by camera arabidopsis, application image Processing method extracts the phenotypic parameter of arabidopsis from image, it is achieved the Non-Destructive Testing of arabidopsis growth course.

Background technology

Arabidopsis is a kind of important model plant in botany, genetics, hereditism.Arabidopsis phenotype is ground Study carefully, can illustrate the physiological function of arabidopsis comprehensively, up hill and dale, the particularly mutual relation between its phenotype and its gene, with And the impact that it is grown by different environmental conditions.The detection method of plant phenotypic characteristics includes that destructive measurement, contact are surveyed Amount and Computer Vision Detection method.Destructive measurement, i.e. for a collection of plant, extracts some randomly, with destructive Method measures its parameter.Contact is measured and is i.e. used touch sensor to measure the parameter of plant.Use computer vision technique Measuring i.e. by relevant device, including CCD camera, light source etc., it is thus achieved that the spectrum picture of measurand, utilization is correlated with Image is processed by software, algorithm, it is thus achieved that required data, thus obtains the phenotypic parameter of plant.Existing research work Realize individual blade is analyzed or the analysis of other crop mainly by computer vision technique.Li Xinguo etc. utilize and sweep Retouch instrument and obtain the image of rape leaf, and utilize Photoshop software to obtain blade pixel count, obtain blade by resolution Area (Li Xinguo, Cai Shengzhong, Li Shaopeng etc. Applied Digital image technique measures avocado leaf area [J]. tropical agricultural science, 2009,29 (2): 10-13.).Han Dianyuan etc. utilize color to carry out the calculation split for the blade proposition under white background is a kind of Method, so utilize reference rectangular slab in background calculate blade area (Han Dianyuan, gold zone deep pool, Fu Hui etc. based on color channel The plant leaf area of similarity image dividing method calculates [J]. Transactions of the Chinese Society of Agricultural Engineering, and 2012,28 (6): 179-183.).Lee is few Elder brothers etc. utilize image technique that Semen Maydis and Semen Tritici aestivi carry out image acquisition, and extract relevant parameter (Li Shaokun, Zhang Xian. crop plant type The research [J] of information multi-media image processing techniques. Acta Agronomica Sinica, 1998,24 (3): 265-271).Li Changying etc. utilize calculating Machine vision technique carries out non-destructive monitoring to hothouse plants growth, obtains the formalness feature of plant, including top projected leaf area With plant height (Li Changying, Teng Guanghui, Zhao Chunjiang etc. utilize computer vision technique realize to hothouse plants growth non-destructive monitoring [J]. Transactions of the Chinese Society of Agricultural Engineering, 2003,19 (3): 140-143.).

In sum, there is following defect in existing research:

1, use destructive measuring method that plant can be caused damage, and plant can not be grown into Line Continuity Measure.

2, use sensor measurement, directly contact with plant, plant can be produced certain impact more or less, And its cost is high, development difficulty is the most relatively large.

3, existing computer vision technique is concentrated mainly on the Phenotypic examination of individual blade or other crop, and to plan The Phenotypic examination of south mustard relies primarily on artificial realization, and workload is big, inefficient.

At present around the COMPUTER DETECTION research of arabidopsis phenotype, document is at home and abroad rarely had to report.

Summary of the invention

The technical problem to be solved in the present invention is: how for the feature of arabidopsis, utilize computer vision technique to plan Mustard carries out Non-Destructive Testing in south, extracts the phenotypic parameter in its growth course.Use normalized rgb value linear arabidopsis image Combination carries out Threshold segmentation, and by again judging the details of plant regional, removes unnecessary noise, it is achieved thereby that not In the case of affecting arabidopsis normal growth, arabidopsis is split from complicated natural growing environment.It is partitioned into plant After, then extract the gross area, blade number and three phenotypic parameters of symmetry.These phenotypic parameters both can describe plan quantitatively The growing state of south mustard, it is also possible to for the research of arabidopsis gene function, i.e. describe different genes by these phenotypic parameters Arabidopsis difference on size and regional morphology, such that it is able to infer heterogeneic function and to arabidopsis thaliana Impact.

(1) technical scheme

To achieve these goals, the invention provides arabidopsis Phenotypic examination method based on computer vision, including Following steps:

S1., in the planting pot of arabidopsis, place scaling board, utilize the RGB image of collected by camera arabidopsis.

S2. to gather after image carry out pretreatment, it is achieved image from dynamic(al) correction and demarcation, wherein image rectification be for The distortion of correction chart picture, image calibration is the full-size(d) in order to obtain unit picture element.

S3. pretreated image is split, by arabidopsis and background segment, extract from image.

S4. after being partitioned into arabidopsis image, extract arabidopsis phenotypic parameter, including the gross area, blade number and Symmetry, describes the growth course of arabidopsis by these parameters.

The scaling board gathered in image uses rim black and white gridiron pattern, by square group that 3 × 3 length of sides are 4mm Become.Place it in Arabidopsis plant side, therewith carry out image acquisition.

Step S2 specifically includes following steps:

S2.1 positions black and white gridiron pattern

S2.1.1, according to blue RGB feature, extracts blue border.

S2.1.2 carries out holes filling to the inside of image, then deducts original blue border image, obtains new images.

S2.1.3 carries out opening operation to new images, removes noise；Carry out closed operation again, connect breakpoint, obtain black and white chessboard The region of lattice.

S2.2 Corner Detection

S2.2.1 calculates the first derivative horizontally and vertically gone up of each point in black and white gridiron pattern region, Obtain three width new images: horizontal first derivative square, the product of two first derivatives of quadratic sum of vertical first derivative.

Three width images described in S2.2.1 are filtered by S2.2.2 gaussian filtering, remove noise.

S2.2.3 is formed correlation matrix, calculation criterion function by S2.2.2 gained three width image, it is judged that pixel therein Whether it is angle point.

S2.3 image rectification and image calibration

S2.3.1 obtains tessellated each foursquare summit coordinate in the picture by Corner Detection, and according to it Spatial relation in real world, obtains the transformation matrix of the two.

S2.3.2 asks for the inverse of transformation matrix, acts on image, it is achieved image rectification.

S2.3.3 obtains the tessellated total number of pixels of black and white by angular coordinate, and according to its full-size(d), obtains The full-size(d) of unit picture element.

Step S3 specifically includes following steps:

S3.1 image foremost segment

S3.1.1 is normalized acquisition rgb to the rgb value of each pixel, extracts the chromaticity difference diagram of 3g-2.4r-b, makees with 0 For threshold value, image is carried out binaryzation.

Result is deducted the blue border region originally detected by S3.1.2, it is judged that the ash of the G of obtained foreground area pixel Whether angle value more than 50, the most then retains, otherwise removes.

S3.1.3 extracts the region with most pixel in prospect connected region, is plant regional.

S3.2 removes noise

S3.2.1 carries out opening operation to gained image, and the result obtained will only retain the center of big leaf area and plant Territory, and remove the detail section in image, including the noise around the stem of plant and blade.

S3.2.2 calculates its number of pixels to each connected region of detail section removed, straight less than 12 of number of pixels Connect removal.

Remaining number of pixels is more than or equal to the connected region of 12 by S3.2.3, the new figure individually obtained with opening operation As superposition, then the areal calculated after superposition in image, if areal reduces, illustrate that this connected region is stem, it is necessary to Retain；Otherwise, if the areal after superposition in image increases or constant, then illustrate that this connected region is around blade Noise, it is necessary to remove.

S3.3.4 can retain the stem of plant by described S3.2.2 and S3.2.3 two-step pretreatment, and removes unwanted making an uproar Point, after finally having judged all of detail section, folds the new images that all connected regions retained all obtain with opening operation Add, obtain final plant regional.

Step S4 specifically includes following steps:

The S4.1 plant gross area

S4.1.1 passes through image segmentation step, obtains the number of the pixel of plant regional.

The true area of the unit picture element that number of pixels obtains with S2 is multiplied by S4.1.2, obtains the true total of plant regional Area.

S4.2 blade number

Plant regional is filled with by S4.2.1, by profile lookup algorithm, extracts the outline point of plant.

S4.2.2 calculates each outline point distance to plant regional central point, thus 2 dimension (2D) images are become 1 dimension (1D) signal.

The signal that S4.2.2 is obtained by S4.2.3 carries out Haar wavelet decomposition, searches the positive zero crossing of the 4th layer of wavelet coefficient, I.e. previous point of certain point less than 0 later point more than 0, what these positive zero crossings represented is 1D signal that S4.2.2 obtains Local maximum.For individual blade, the central point of its blade cusp distance plant regional is farthest, i.e. the office of 1D signal Portion's maximum is the cusp of each blade.Therefore the blade number of plant it is achieved with by the local maximum searching 1D signal.

S4.3 symmetry

S4.3.1 extracts abscissa and the vertical coordinate of plant regional, by calculating its respective variance and covariance, obtains Its covariance matrix, calculates the characteristic vector of this covariance matrix, obtains two major axes orientations of plant regional.

Plant regional is overturn respectively by S4.3.2 along the two main shaft, two images upset obtained and artwork As being overlapped, calculate the number of pixels of overlapping plant regional.

Two overlapping number of pixels respectively compared with total number of pixels of plant regional, are obtained both direction by S4.3.3 Symmetry.

(2) beneficial outcomes

The inventive method utilizes computer vision technique to gather the image of arabidopsis, utilizes image processing techniques to achieve plan The Non-Destructive Testing of south mustard phenotype.By the detection tessellated blue border of rim black and white and internal with Corner Detection Algorithm detection Angle point, finally achieve image from dynamic(al) correction and demarcation.Normalized rgb value linear combination is used to carry out threshold arabidopsis Value segmentation, and by again judging the details of plant regional, remove unnecessary noise, it is achieved thereby that do not affecting arabidopsis In the case of normal growth, arabidopsis is split from complicated natural growing environment.After being partitioned into plant, then extract The gross area, blade number and three phenotypic parameters of symmetry.Observing relative to Traditional Man and measure, the method has in efficiency Bigger raising.

Accompanying drawing explanation

By the detailed description non-limiting example made with reference to the following drawings of reading, the further feature of the present invention, Purpose and advantage will become more apparent upon:

Fig. 1 is the flow process of the arabidopsis Phenotypic examination method based on computer vision according to one embodiment of the invention Figure；

The collection transposition for gathering image used in the method that Fig. 2 provides for the present invention；

The image generated in the preprocess method processing procedure that Fig. 3 provides for the present invention, wherein, subgraph (1) be according to The RGB image of the arabidopsis that inventive method gathers, (2) are the black and white gridiron pattern region generated after pretreatment, and (3) are for passing through The Corner Detection image generated after pretreatment, (4) are the image after the correction generated after pretreatment；

The image generated in the image partition method processing procedure that Fig. 4 provides for the present invention, wherein, subgraph (1) is for passing through The bianry image generated after image foremost segment step, (2) are the figure of the final arabidopsis generated after noise removal step Picture；

The image generated in the phenotype extracting method processing procedure that Fig. 5 provides for the present invention, wherein, subgraph (1) is for passing through The outline of the plant regional that profile lookup algorithm extracts, (2) are each outline point distance to central point, and (3) are for being somebody's turn to do The 4th layer of wavelet coefficient that distance signal obtains after Haar wavelet decomposition, (4) are the image of the blade cusp detected.

Detailed description of the invention

The arabidopsis Phenotypic examination method based on computer vision that the present invention proposes, the most specifically Bright as follows.Following example will assist in those skilled in the art and are further appreciated by the present invention, but limit the most in any form The present invention.It should be pointed out that, to those skilled in the art, without departing from the inventive concept of the premise, also Some deformation and change can be made.These broadly fall into protection scope of the present invention.

For the phenotypic parameter of rapid extraction arabidopsis, the present invention proposes arabidopsis Phenotypic examination based on computer vision Method.After the method carries out pretreatment to the image gathered, from complicated background environment, it is partitioned into arabidopsis, basis at this On the gross area, blade number and the symmetry of arabidopsis are extracted, the phenotypic parameter that improve arabidopsis obtains efficiency.

As it is shown in figure 1, include step according to the arabidopsis Phenotypic examination method of one embodiment of the invention:

S1. in the planting pot of arabidopsis, place scaling board, utilize CCD camera to gather the RGB image of arabidopsis；

In this example, the harvester that can use Fig. 2 carries out the collection of image.This harvester includes: CCD camera 1, support 2, illuminator 3, arabidopsis 4 and scaling board 5.Scaling board 5 is placed on the side of arabidopsis 4.Scaling board 5 uses indigo plant Frame black and white gridiron pattern, is that the square of 4mm forms by 3 × 3 length of sides.

S2. the image after gathering is carried out pretreatment, specifically includes following sub-step:

S2.1 positions black and white gridiron pattern

S2.1.1 is according to blue RGB feature, and the gray value of R and G is both less than B, and the gray value of B is more than 150, from image In be partitioned into blue border.

S2.1.2 carries out holes filling to the inside of image, and its result deducts original blue border image, is newly schemed Picture.

S2.1.3 carries out opening operation to new images, removes noise；Carry out closed operation again, connect breakpoint, obtain black and white chessboard The region of lattice.

S2.2 Corner Detection

S.2.2.1 the single order horizontally and vertically gone up calculating each point in black and white gridiron pattern region is led Number.Use following Prewitt template to calculate, be i.e. approximately horizontal direction by the 3rd row in 3 × 3 regions and the difference of first row Derivative, be approximately the derivative of vertical direction by the difference of the third line and the first row.By the two subtemplate and image convolution, obtain Two matrixes identical with image size, are designated as I_x, I_y, and then calculate three width new images: I_x ², I_y ²And I_xI_y。

Prewitt template

S.2.2.2 can be disturbed by noise in view of image, use the Gauss window of 101 × 101 that three width images are entered Row filtering, removes noise.

S.2.2.3 correlation matrix M is formed by three width new images:

$M=\left[\begin{array}{cc}{I}_x^2& I_x{I}_y\\ I_x{I}_y& I_y^2\end{array}\right]$

Utilize this matrix calculus criterion function R:

R=det(M)-k·(trace(M))²

Wherein k typically takes 0.04.

For each pixel, a R value can be obtained, if the R value of certain point is more than 0.01R_max, and it is 3 × 3 The local maximum of neighborhood, then it will be judged as angle point.

S2.3 image rectification and image calibration

S.2.3.1 tessellated each foursquare summit coordinate in the picture is obtained by Corner Detection, in true generation In boundary, these summits form each square, according to its spatial relation, can arrange the coordinate on these summits.Order is wherein Certain point coordinate in the picture is (x ', y ')^T, coordinate in space be (x, y)^T, its homogeneous coordinates are respectively (x₁', X₂', x₃’)^T(x₁, y, 1)^T, there is therebetween Projective Distortion conversion, transformation matrix is linear about homogeneous three-dimensional coordinate Conversion H, is expressed as:

$\left[\begin{array}{c}{x}_1^{\′}\\ x_2^{\′}\\ x_3^{\′}\end{array}\right]=\left[\begin{array}{ccc}{h}_{11}& h_{12}& h_{13}\\ h_{21}& h_{22}& h_{23}\\ h_{31}& h_{32}& h_{33}\end{array}\right]\left[\begin{array}{c}x\\ y\\ 1\end{array}\right]$

Because the inhomogeneous coordinate of image can be expressed as by homogeneous coordinates:

$x^{\′}=\frac{x_1^{\′}}{x_3^{\′}}=\frac{h_{11}x+h_{12}y+h_{13}}{h_{31}x+h_{32}y+h_{33}}$

$y^{\′}=\frac{x_2^{\′}}{x_3^{\′}}=\frac{h_{21}x+h_{22}y+h_{23}}{h_{31}x+h_{32}y+h_{33}}$

So each group of match point can get two equation below:

(h₃₁x+h₃₂y+h₃₃) x '=h₁₁x+h₂y+h₁₃

(h₃₁x+h₃₂y+h₃₃)y′=h₂₁x+h₂₂y+h₂₃

Transformation matrix H can be obtained with at least four pairs of match points.

S.2.3.2 ask for the inverse of transformation matrix H, remake for image, it is achieved the correction of image.

S2.3.3 calculates total number of pixels N in black and white gridiron pattern region by angular coordinate_b, and according to its true chi Very little, i.e. 9 × (4mm)²=144mm², obtain the full-size(d) of unit picture element, it may be assumed that

$A_{\mathrm{dz}}=\frac{144}{N_b}$

Pretreated image is split by S3, is extracted by arabidopsis from image, specifically includes following sub-step Rapid:

S3.1 image foremost segment

S3.1.1 is normalized acquisition rgb to the rgb value of each pixel, and method is as follows:

$\left\{\begin{array}{c}r=\frac{R}{R+G+B}\\ g=\frac{G}{R+G+B}\\ b=\frac{B}{R+G+B}\end{array}\right.$

Extract the chromaticity difference diagram of 3g-2.4r-b, as threshold value, image is carried out binaryzation with 0, even

I=3g-2.4r-b≥0

Then this pixel is judged as the pixel of prospect plant, is otherwise background pixel.

Result is deducted the blue border region originally detected by S3.1.2.Judge the G's of the pixel of obtained foreground area Whether gray value more than 50, the most then retains this pixel, otherwise this pixel is judged as background pixel, removes from region.

S3.1.3 extracts the region with most pixel in prospect connected region, is plant regional.

S3.2 removes noise

S3.2.1 carries out opening operation to gained image by the disk template of 7 × 7, and the result obtained will only retain big vane region Territory and the central area of plant, and remove the detail section in image, including the noise around the stem of plant and blade.

S3.2.2 calculates its number of pixels to each connected region of detail section removed, straight less than 12 of number of pixels Connect removal.

The result that each details area remaining obtains with the first step respectively is overlapped by S3.2.3.If stem, its By the blade in Contiguous graphics and plant central area so that the connected region number in last image tails off.If attachment Noise around blade, then, after superposition, the connected region number in image becomes many or constant.By connecting after judging superposition The number in region, removes each details or retains.

S3.3.4 can retain the stem of plant by first two steps, and removes unwanted noise, finally will be by judging after The result that the details retained obtains with the first step superposes, and obtains last plant regional.

S4., after being partitioned into arabidopsis, the phenotypic parameter of arabidopsis is extracted, including the gross area, blade number and symmetry Property, specifically includes following sub-step:

The S4.1 plant gross area

S4.1.1 passes through image segmentation step, obtains number N of the pixel of plant regional_p。

The true area of the unit picture element that number of pixels obtains with the image calibration step in S2 is multiplied by S4.1.2, obtains The true gross area of plant regional, it may be assumed that

A_z=A_dzN_p

S4.2 blade number

Plant regional is filled with by S4.2.1, by profile lookup algorithm, extracts the outline point of plant.

S4.2.2 calculates each outline point distance to plant regional central point.Make certain outline point in the picture Coordinate is (x_i, y_i), central point coordinate in the picture is (x_o, y_o), distance therebetween is:

$D_i=\sqrt{{(x_i-x_o)}^2+{(y_i-y_o)}^2}$

By calculating distance D of each point_i, obtain a 1D signal S, thus 2 dimension (2D) images become 1 dimension (1D) letter Number

S4.2.3 carries out Haar wavelet decomposition to signal S, extracts the 4th layer of wavelet coefficient, it is thus achieved that new array S_C.Right S_CMove to right one respectively and move to left one, obtaining S_C1And S_C2.Search and meet the point of following condition:

(S_C=0I S_C2>0)∪(S_C1<0I S_c<0IS_C2>0)

The point meeting above-mentioned condition is the positive zero crossing of S signal, i.e. previous point of certain point less than 0 later point More than 0, what these positive zero crossings represented is the local maximum of 1D signal S.For individual blade, its blade cusp distance The central point of plant regional is farthest, i.e. the local maximum of 1D signal is the cusp of each blade.Therefore by searching 1D signal Local maximum is achieved with the blade number of plant.

S4.3 symmetry

S4.3.1 extracts abscissa and the vertical coordinate of plant regional, by calculating its respective variance and covariance, obtains Its covariance matrix.

$M=\left[\begin{array}{cc}{\mathrm{\&sigma;}}_{\mathrm{xx}}& {\mathrm{\&sigma;}}_{\mathrm{xy}}\\ {\mathrm{\&sigma;}}_{\mathrm{xy}}& {\mathrm{\&sigma;}}_{\mathrm{yy}}\end{array}\right]$

Calculate the characteristic vector of this covariance matrix, obtain two major axes orientations of plant regional.

Plant regional is overturn respectively by S4.3.2 along the two main shaft, two images upset obtained and artwork As being overlapped, calculate the number of pixels of overlapping plant regional.

Two overlapping number of pixels respectively compared with total number of pixels of plant regional, are obtained both direction by S4.3.3 Symmetry.

The method further illustrating the present invention below in conjunction with examples of implementation, the method comprises the following steps:

S1, in the planting pot of arabidopsis, places scaling board, utilizes the RGB image of collected by camera arabidopsis.Scaling board is adopted Be rim black and white gridiron pattern, be made up of the square that 3 × 3 length of sides are 4mm.Place it in arabidopsis side, with one Rise and carry out image acquisition, shown in the picture of collection such as Fig. 3 (1).

S2 to gather after image carry out pretreatment, it is achieved image from dynamic(al) correction and demarcation.

S2.1 positions black and white gridiron pattern

S2.1.1 is according to blue RGB feature, and the gray value of R and G is both less than B, and the gray value of B is more than 150, from image In be partitioned into blue border.

S2.1.2 carries out holes filling to the inside of image, and its result deducts original blue border image, is newly schemed Picture.

S2.1.3 carries out opening operation to new images, removes noise；Carry out closed operation again, connect breakpoint, obtain black and white chessboard The region of lattice, as shown in Fig. 3 (2).

S2.2 Corner Detection

S2.2.1 Prewitt template and image convolution, calculate the level side that each in black and white gridiron pattern region is put To with the first derivative in vertical direction, obtain three width new images: horizontal first derivative square, vertical first derivative square Product with two first derivatives.

S2.2.2 uses the Gauss window of 101 × 101 to be filtered three width images, removes noise.

S2.2.3 is formed correlation matrix, calculation criterion function by above-mentioned three width images, finds out angle point.As shown in Fig. 3 (3), Each angle point marks with red point.

S2.3 image rectification and image calibration

S2.3.1 obtains tessellated 16 foursquare summits according to from left to right by Corner Detection, from top to bottom Order coordinate in the picture is as follows:

(163,370)(224,362)(286,354)(347,347)

(171,430)(233,423)(294,415)(356,407)

(179,491)(241,484)(302,476)(346,468)

(187,553)(249,545)(311,537)(372,529)

It is strict square in real world according to it, by each coordinate convention put is:

(164,347)(225,347)(286,347)(347,347)

(164,408)(225,408)(286,408)(347,408)

(164,469)(225,469)(286,469)(347,469)

(164,530)(225,530)(286,530)(347,530)

Calculating these two groups of transformation of coordinates matrixes is:

$\left[\begin{array}{ccc}0.9820& 0.1202& -1.2519e-05\\ -0.1232& 1.0126& 3.8417e-05\\ 51.0888& -43.5533& 0.9977\end{array}\right]$

The inverse of this transformation matrix of S2.3.2 is:

$\left[\begin{array}{ccc}1.0026& -0.1183& 0.0000\\ 0.1237& 0.9713& 0.0000\\ -45.9428& 48.4665& 1.0000\end{array}\right]$

Acted on image, it is achieved image rectification, recovered the metric characteristic of image, shown in result such as Fig. 3 (4).

It is 35036 that S2.3.3 obtains the tessellated total number of pixels of black and white by angular coordinate, and according to its true chi Very little, i.e. 9 × (4mm)²=144mm², obtain the full-size(d) of unit picture element, it may be assumed that

$A_{\mathrm{dz}}=\frac{144}{35036}$

Pretreated image is split by S3, is extracted by arabidopsis from image, specifically includes following sub-step Rapid:

S3.1 image foremost segment

S3.1.1 is normalized acquisition rgb to the rgb value of each pixel, extracts the chromaticity difference diagram of 3g-2.4r-b, makees with 0 For threshold value, image is carried out binaryzation.

Result is deducted the blue border region originally detected by S3.1.2.Judge the G's of the pixel of obtained foreground area Whether gray value more than 50, the most then retains this pixel, otherwise this pixel is judged as background pixel, removes from region.

S3.1.3 extracts the region with most pixel in prospect connected region, is plant regional, result such as Fig. 4 (1) shown in.

S3.2 removes noise

S3.2.1 carries out opening operation to gained image by the disk template of 7 × 7, and the result obtained will only retain big vane region Territory and the central area of plant, and remove the detail section in image, including the noise around the stem of plant and blade.

S3.2.2 calculates its number of pixels to each connected region of detail section removed, straight less than 12 of number of pixels Connect removal.

The result that each details area remaining obtains with the first step respectively is overlapped by S3.2.3.If stem, its By the blade in Contiguous graphics and plant central area so that the connected region number in last image tails off.If attachment Noise around blade, then, after superposition, the connected region number in image becomes many or constant.By connecting after judging superposition The number in region, removes each details or retains.

S3.3.4 can retain the stem of plant by first two steps, and removes unwanted noise, finally will be by judging after The result that the details retained obtains with the first step superposes, and obtains last plant regional, shown in result such as Fig. 4 (2).

S4., after being partitioned into arabidopsis, the phenotypic parameter of arabidopsis is extracted, including the gross area, blade number and symmetry Property, specifically includes following sub-step:

The S4.1 plant gross area

S4.1.1 passes through image segmentation step, and the number of the pixel obtaining plant regional is 35430.

The true area of the unit picture element that number of pixels obtains with S2 is multiplied by S4.1.2, obtains the true total of plant regional Area, it may be assumed that

$A_{\mathrm{dz}}=\frac{144}{35036}\&times;35430={145.6\mathrm{mm}}^2$

S4.2 blade number

Plant regional is filled with by S4.2.1, by profile lookup algorithm, extracts the outline point of plant, such as Fig. 5 (1) Shown in.

S4.2.2 calculates each outline point distance to plant regional central point, as shown in Fig. 5 (2).

The signal that S4.2.2 is obtained by S4.2.3 carries out Haar wavelet decomposition, searches the zero crossing of the 4th layer of wavelet coefficient, as Shown in Fig. 5 (3), thus obtain the local maximum of distance signal, thus obtain the cusp of each blade, it is thus achieved that the blade of plant Number is 8, as shown in Fig. 5 (4).

S4.3 symmetry

S4.3.1 extracts abscissa and the vertical coordinate of plant regional, by calculating its respective variance and covariance, obtains Its covariance matrix.

$M=\left[\begin{array}{cc}4632.3& 257.8\\ 257.8& 3325.1\end{array}\right]$

It is calculated two characteristic vectors of this covariance matrix respectively:

$V_1=\left[\begin{array}{c}0.1867\\ -0.9824\end{array}\right]$

$V_2=\left[\begin{array}{c}-0.9824\\ -0.1867\end{array}\right]$

Thus obtain two major axes orientations of plant regional respectively:

${\mathrm{\&theta;}}_1=\arctan \left(\frac{-0.9824}{0.1867}\right)=-79.23$

${\mathrm{\&theta;}}_2=\arctan \left(\frac{-0.1867}{-0.9824}\right)=10.76$

Plant regional is overturn respectively by S4.3.2 along the two main shaft, two images upset obtained and artwork As being overlapped, the number of pixels of the plant regional obtaining overlap is 24022 and 22734 respectively.

Two overlapping number of pixels respectively compared with total number of pixels of plant regional, are obtained both direction by S4.3.3 Symmetry respectively:

$P_1=\frac{24022}{35430}=0.6780$

$P_1=\frac{22734}{35430}=0.6417$

Claims (4)

1. an arabidopsis Phenotypic examination method, comprises the following steps:

S1., in the planting pot of arabidopsis, place scaling board, utilize the RGB image of collected by camera arabidopsis；

S2. the image after gathering is carried out pretreatment, it is achieved image from dynamic(al) correction and demarcation, wherein image calibration is precisely in order to school The distortion of positive image, image calibration is the full-size(d) in order to obtain unit picture element；

Step S2 includes positioning black and white gridiron pattern, Corner Detection, image rectification and image calibration；

Described location black and white gridiron pattern comprises the following steps: according to the blue component in RGB feature, extracts blue border；To figure The inside of picture carries out holes filling, then deducts original blue border image, obtains new images；New images is carried out opening operation, Remove noise；Carry out closed operation again, connect breakpoint, obtain the tessellated region of black and white；

Described Corner Detection comprises the following steps: calculate horizontal direction and Vertical Square that each in black and white gridiron pattern region is put First derivative upwards, obtains three width new images: horizontal first derivative square, two single orders of the quadratic sum of vertical first derivative The product of derivative；With gaussian filtering, three width images are filtered, remove noise；Correlation matrix is formed by above-mentioned three width images, Calculation criterion function, it is judged that whether pixel therein is angle point；

Described image rectification and image calibration comprise the following steps: obtain tessellated each foursquare top by Corner Detection Point coordinate in the picture, and according to its spatial relation in real world, obtain the transformation matrix of the two；Ask for conversion Inverse of a matrix, acts on image, it is achieved image rectification；The tessellated total number of pixels of black and white, and root is obtained by angular coordinate According to its full-size(d), obtain the full-size(d) of unit picture element；

S3. pretreated image is split, by arabidopsis and background segment, extract from image；

Step S3 includes image foremost segment and removes noise；

Described image foremost segment comprises the following steps: the rgb value of each pixel is normalized acquisition rgb, extracts 3g- The chromaticity difference diagram of 2.4r-b, carries out binaryzation as threshold value to image with 0；Result is deducted the blue border district originally detected Territory, it is judged that whether the gray value of the G of obtained foreground area pixel more than 50, the most then retains, otherwise remove；Extraction prospect is even The region with most pixel in logical region, is plant regional；

Described removal noise comprises the following steps: gained image is carried out opening operation, obtains new image, and it will only retain great Ye Panel region and the central area of plant, and remove the detail section in image, including the noise around the stem of plant and blade；Right Each connected region of the detail section removed calculates its number of pixels, number of pixels directly removing less than 12, number of pixels Connected region more than or equal to 12, the new images individually obtained with opening operation superposes, then the district calculated after superposition in image Territory number, if areal reduces, illustrates that this connected region is stem, it is necessary to retain, otherwise, if the district after superposition in image Territory number increases or constant, then illustrate that this connected region is the noise around blade, it is necessary to remove；Judge all of details After part, the new images that all connected regions retained all obtain with opening operation is superposed, obtains final plant regional；

S4., after being partitioned into arabidopsis image, the phenotypic parameter extracting arabidopsis describes the growth of arabidopsis by these parameters Journey；Wherein arabidopsis phenotypic parameter is the arabidopsis thaliana gross area, Arabidopsis leaf number and arabidopsis symmetry.

Arabidopsis Phenotypic examination method the most according to claim 1, the wherein said arabidopsis thaliana gross area is by figure As segmentation step, obtain the number of the pixel of plant regional；The list that number of pixels is obtained with the image calibration step in S2 The true area of position pixel is multiplied, and obtains the true gross area of plant regional.

Arabidopsis Phenotypic examination method the most according to claim 1, wherein said Arabidopsis leaf number is by plant Region is filled with, and by profile lookup algorithm, extracts the outline point of plant；Calculate each outline point in plant regional The distance of heart point, thus 2 dimension (2D) images are become 1 dimension (1D) signal；1D signal is carried out Haar wavelet decomposition, searches the 4th Later point is more than 0 less than 0 for the positive zero crossing of layer wavelet coefficient, the i.e. previous point of certain point, and these positive zero crossings represent Be the local maximum of this 1D signal；For individual blade, the central point of its blade cusp distance plant regional is farthest, The local maximum of i.e. 1D signal is the cusp of each blade；Therefore it is achieved with plant by the local maximum searching 1D signal Blade number.

Arabidopsis Phenotypic examination method the most according to claim 1, wherein said arabidopsis symmetry is by extracting plant The abscissa in region and vertical coordinate, by calculating its respective variance and covariance, obtain its covariance matrix, calculate this association side The characteristic vector of difference matrix, obtains two major axes orientations of plant regional；Plant regional is carried out respectively along the two main shaft Upset, two images upset obtained are overlapped with original image, calculate the number of pixels of overlapping plant regional；By two Overlapping number of pixels, respectively compared with total number of pixels of plant regional, obtains the symmetry of both direction.