CN103810522A - Counting method and device for corn ear grains - Google Patents

Abstract

The invention provides a counting method and a counting device for corn ear grains. The method comprises the following steps: step S1, obtaining an estimation model for the corn ear grains, wherein the estimation model for the corn ear grains is used for showing a function relation between the number of the ear grains on the single surface of corn and the number of all ear grains; step 2, obtaining the single sided image of the corn, and preprocessing the single sided image to obtain a corresponding gray level image; step S3, performing multi-threshold segmentation on the gray level image obtained in the step S2, so as to obtain a corresponding binary image; step S4, counting communicating areas with the same gray values, and obtaining the number of the communication areas corresponding to the grey values for showing the corn grains; and step S5, determining all ear grains according to the number obtained in the step S4 and the corn grain estimation model in the step S1. By adopting the counting method for the corn ear grains, the corn ear grains can be quickly and conveniently counted.

Description

A kind of corn ear seed method of counting and device

Technical field

The present invention relates to image processing techniques, relate in particular to a kind of corn ear seed method of counting and device.

Background technology

Corn seed selection is the heterozygosity of utilizing corn cross-pollination and hereditary basis, and a large amount of self-mating systems and assorted combo are closed and carried out cross experiment, and the cenospecies of finally selecting indivedual excellences puts into production use.Wherein the seed-producing rate of fruit ear is the important indicator of passing judgment on corn variety quality.

At present China's corn ear seed counting work is mainly take manually as main.This mode not only workload is very huge, and reading out data error is large, and the data that cannot repeat having read read again, goes wrong and is difficult to verify that it corrects errors.Developed country's researchs such as America and Europe are used species test instrument.But this quasi-instrument selling at exorbitant prices, volume weight is large, is difficult for flowing using, and is not suitable in production reality, promoting and using.

Therefore develop a kind of automated computer and assist to improve corn ear seed counting efficiency, improve method of counting, examination and the popularization of accelerating kind are very necessary.

Summary of the invention

The invention provides a kind of corn ear seed method of counting and device, can realize the fast and convenient counting of corn ear seed,

The invention provides a kind of corn ear seed method of counting, the method comprises:

A kind of corn ear seed method of counting, is characterized in that, comprising:

Step S1, obtains corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Step S2, obtains the one-sided image of corn, and described one-sided image is carried out to image pre-service, obtains corresponding gray level image;

Step S3, carries out many Threshold segmentations to the gray level image obtaining in step S2, obtains corresponding bianry image;

Step S4, counts the connected region with same grayscale value, obtains the number with the connected region that the gray-scale value for representing iblet is corresponding;

Step S5, determines whole fruit ear kernal numbers according to the iblet appraising model in the number getting in step S4 and step S1.

Preferably, described step S1 specifically comprises:

Obtain following formula as corn ear seed appraising model:

y=ax+b；

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

Preferably, described step S2 comprises:

Step S21, carries out medium filtering processing to described one-sided image;

Step S22, carries out Wallis filtering processing to the image obtaining in step S21;

Step S23, carries out histogram enhancement processing to the image obtaining in step S22, obtains the gray level image that described one-sided image is corresponding.

Preferably, described step S3, specifically comprises:

Take many threshold segmentation methods that RBGM and OTSU combine to carry out to the gray level image obtaining in step S2 the bianry image that corn kernel is cut apart and then obtained to image.

Preferably, described in take many threshold segmentation methods that RBGM and OTSU combine to carry out to corn ear seed gray level image the bianry image that corn kernel is cut apart and then obtained to image, specifically comprise:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold for making

get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\η}}_m\frac{\∂{\mathrm{\σ}}_B^2}{\∂t_i},i=\mathrm{1,2},...,N;$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\σ}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\Σ}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\Σ}}}{\mathrm{\ω}}_k{\mathrm{\ω}}_j{({\mathrm{\μ}}_k-{\mathrm{\μ}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart;

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

The invention provides a kind of corn ear seed counting assembly, comprising:

MBM, for obtaining corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Image pretreatment module, for obtaining the one-sided image of corn, and carries out image pre-service to described one-sided image, obtains corresponding gray level image;

Image is cut apart module, carries out many Threshold segmentations for the gray level image that described image pretreatment module is obtained, and obtains corresponding bianry image;

Counting module, for counting, obtain the number with the connected region that the gray-scale value for representing iblet is corresponding to the connected region with same grayscale value;

Computing module, the iblet appraising model obtaining according to the number getting in described counting module and described MBM is determined whole fruit ear kernal numbers.

Preferably, described MBM is specifically for obtaining following formula as corn ear seed appraising model:

y=ax+b；

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

Preferably, described image pretreatment module comprises:

First processes submodule, for described one-sided image is carried out to medium filtering processing;

Second processes submodule, carries out Wallis filtering processing for the image that the first processing submodule is obtained;

The 3rd processes submodule, carries out histogram enhancement processing for the image that the second processing submodule is obtained, and obtains the gray level image that described one-sided image is corresponding.

Preferably, described image is cut apart module, specifically for taking the gray level image that many threshold segmentation methods that RBGM and OTSU combine obtain described image pretreatment module to carry out the bianry image that corn kernel is cut apart and then obtained to image.

Preferably, described image is cut apart gray level image that module obtains described image pretreatment module for many threshold segmentation methods of taking RBGM and OTSU and combining and is carried out image when cutting apart and then obtaining the bianry image of corn kernel, specifically for carrying out:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold

for making

get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\η}}_m\frac{\∂{\mathrm{\σ}}_B^2}{\∂t_i},i=\mathrm{1,2},...,N;$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\σ}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\Σ}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\Σ}}}{\mathrm{\ω}}_k{\mathrm{\ω}}_j{({\mathrm{\μ}}_k-{\mathrm{\μ}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart;

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

Adopt corn ear seed method of counting provided by the invention, can realize the fast and convenient counting of corn ear seed.

Accompanying drawing explanation

The schematic flow sheet of a kind of corn ear seed method of counting that Fig. 1 provides for the embodiment of the present invention;

The schematic flow sheet of gray scale processing in the corn ear seed method of counting that Fig. 2 provides for the embodiment of the present invention;

The schematic flow sheet that in the corn ear seed method of counting that Fig. 3 provides for the embodiment of the present invention, image is cut apart;

One side corn map picture after the corn ear seed method of counting that Fig. 4 a provides for the use embodiment of the present invention identifies corn ear;

Local one side corn map picture after the corn ear seed method of counting that Fig. 4 b provides for the use embodiment of the present invention identifies corn ear;

The structural representation of the corn ear seed counting assembly that Fig. 5 provides for the embodiment of the present invention.

Embodiment

Below in conjunction with drawings and Examples, the specific embodiment of the present invention is further described.Following examples are only for technical scheme of the present invention is more clearly described, and can not limit the scope of the invention with this.

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail.Following examples are used for illustrating the present invention, but do not limit the scope of the invention.

Embodiment mono-

A kind of corn ear seed counting is provided, and as shown in Figure 1, the method comprises:

Step S1, obtains corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Step S2, obtains the one-sided image of corn, and described one-sided image is carried out to image pre-service, obtains corresponding gray level image;

Step S3, carries out many Threshold segmentations to the gray level image obtaining in step S2, obtains corresponding bianry image;

Step S4, counts the connected region with same grayscale value, obtains the number with the connected region that the gray-scale value for representing iblet is corresponding;

Step S5, determines whole fruit ear kernal numbers according to the iblet appraising model in the number getting in step S4 and step S1.

Adopt corn ear seed method of counting provided by the invention, can realize the fast and convenient counting of corn ear seed.

Realizing in process of the present invention, present inventor finds: (1) corn ear seed line number is substantially fixing, is generally 14 row or 16 row; (2) when corn image sampling, use erect image projecting method, corn kernel accounts for corn kernel region at image visible part and fixes, on this basis, in the preferred embodiment of the invention, described step S1 can be specially: obtain following formula (1) as corn ear seed appraising model:

y=ax+b； （1）

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

Here a, b can obtain by training, by part corn ear is carried out to artificial counting, and a, b value are carried out to assignment.It is pointed out that the training pattern of setting up can be used for many times for step S2-S5, is convenient to realize the quick counter to other corn ear seeds here.

Because step S4 is the result of the one-sided image to corn, the seed therefore obtaining is also one side, is equivalent to the x in above-mentioned relation formula, now x is brought in above-mentioned relational expression, can obtain total number of corn ear seed.

Certainly in practical application, those skilled in the art also can use other mathematical model, specifically use which kind of model not affect protection scope of the present invention.

Preferably, in above-mentioned step S2, one-sided image being carried out to the pretreated method of image can comprise as shown in Figure 2:

Step S21, carries out medium filtering processing to described one-sided image;

Step S22, carries out Wallis filtering processing to the image obtaining in step S21;

Step S23, carries out histogram enhancement processing to the image obtaining in step S22, obtains the gray level image that described one-sided image is corresponding.

In practical application, before step S21, also can get corn fruit fruit ear seed gray level image f (x, y), wherein f (x, y) is the gray-scale value that (x, y) pixel is corresponding.

Wherein, the formula that the medium filtering in step S21 adopts in processing is following formula (2):

$f^{\text{'}}(x_0,y_0)={\left[\underset{(x_i,y_i)\∈S}{\mathrm{Sort}}{\mathrm{\ω}}_{\mathrm{ij}}f(x_i,y_i)\right]}_{\frac{\left|S\right|+1}{2}}---\left(2\right)$

Wherein S is pixel (x ₀, y ₀) neighborhood set wherein comprise (x ₀, y ₀), (x, y) represents the arbitrary element in S, f (x, y) represents the gray-scale value of (x, y) point, | S| represents the number of element in S set, and Sort represents sequence, and ω represents the weights that add while calculating intermediate pixel value.

In step S22, to process the formula adopting be following formula (3) in Wallis filtering:

$g(i,j)=[\mathrm{\α}{m}_d+(1-\mathrm{\α})\stackrel{\‾}{f}(i,j)]+[f(i,j)-\stackrel{\‾}{f}(i,j)]\frac{A{\mathrm{\σ}}_d}{\mathrm{A\σ}(i,j)+{\mathrm{\σ}}_d}---\left(3\right)$

Wherein [f (i, j)] _{m × N}represent original image, [g (i, j)] _{m × N}represent the image after strengthening,

represent local mean value, m _drepresent mean value, σ _drepresent standard deviation, A represents gain coefficient, and α controls the proportionality constant that strengthens edge and background composition in image.

In step S23, to process the formula adopting be following formula (4) to histogram enhancement

z＝G ^-1[T(r)] （4）

Pixel grayscale before r presentation video strengthens, transforming function transformation function s=T (r) represents the grey scale pixel value after enhancing, inverse transform function z=G ^-1(s) represent desired new gray level.

Preferably, described step S3, specifically comprises:

Take many threshold segmentation methods that RBGM and OTSU combine to carry out to the gray level image obtaining in step S2 the bianry image that corn kernel is cut apart and then obtained to image.

Concrete, step S3 can comprise each step as shown in Figure 3:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula (5) is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold

for making

get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\η}}_m\frac{\∂{\mathrm{\σ}}_B^2}{\∂t_i},i=\mathrm{1,2},...,N;---\left(5\right)$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\σ}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\Σ}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\Σ}}}{\mathrm{\ω}}_k{\mathrm{\ω}}_j{({\mathrm{\μ}}_k-{\mathrm{\μ}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart.

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

The principle of above-mentioned steps S31 is:

Suppose that population variance is labeled as σ for n class segmentation problem ², all kinds of probability, average, variances are labeled as respectively ω _k, μ _k,

(k=0,1 ..., n-1), the composite marking of each segmentation threshold is t={t _k| k=1,2 ..., n-1}, has:

wherein k=0,1 ...., n-1(6)

t ₀=0, t _n=L is histogram border (7)

wherein 1 < t _k< L (k=1,2 ..., n-1) be each segmentation threshold, (8)

Each class internal variance of many threshold values is:

${\mathrm{\&sigma;}}_w^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&sigma;}}_k^2---\left(9\right)$

Be similar to single threshold problem, set up the equation of constraint of many Thresholds:

${\mathrm{\&sigma;}}^2=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}{(i-\mathrm{\&mu;})}^2{P}_i=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}{i}^2{P}_i-2\mathrm{\&mu;}\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}i{P}_i+{\mathrm{\&mu;}}^2\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}{P}_i=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}{i}^2{P}_i-{\mathrm{\&mu;}}^2;$

Can draw according to above formula ${\mathrm{\&sigma;}}^2-{\mathrm{\&sigma;}}_w^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&mu;}}_k^2+\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&mu;}}_k\underset{j\&NotEqual;k}{\mathrm{\&Sigma;}}{\mathrm{\&omega;}}_j{\mathrm{\&mu;}}_j;$

Derive and can draw through arranging ${\mathrm{\&sigma;}}_B^2={\mathrm{\&sigma;}}^2-{\mathrm{\&sigma;}}_w^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&omega;}}_j{({\mathrm{\&mu;}}_k-{\mathrm{\&mu;}}_j)}^2---\left(10\right);$ Obviously

for with respect to independent variable t _ifunction.

Afterwards by the x in gradient descent method (RBGM) formula (11) _ireplace with t _i, by f(x) replace with

can fall for solution formula (12):

Wherein,

j＝0,1,...,N-1;m＝0,1,...。

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\&eta;}}_m\frac{\&PartialD;{\mathrm{\&sigma;}}_B^2}{\&PartialD;t_i},i=\mathrm{1,2},...,N;---\left(12\right).$

Many threshold segmentation methods that employing RBGM and OTSU combine are for processing in the many threshold segmentation methods of traditional OSTU along with segmentation threshold number is increased to many threshold values from single threshold, the calculated amount that optimal threshold is solved will become the problem of geometric growth, adopt the method can under same precision, greatly reduce separation calculation amount, and then shorten the separation calculation time.

The bianry image of available corn ear seed after step S3, wherein each corn kernel is divided into the region of a connection, now, the connected region with same grayscale value is counted, obtain the number with the connected region that the gray-scale value for representing iblet is corresponding, calculate corn ear seed sum according to the corn ear seed appraising model of setting up in advance afterwards.As Fig. 4 (a) (b) as shown in, in the bianry image of corn ear seed, each corn kernel is divided into the region of a connection, in this region, adjacent pixel has same gray-scale value 0 or 1.Counter C of first initialization before mark, initial value is 1, is used for counting one side corn kernel number, p (x, y) represent pixel is at the pixel value of (x, y) point.From any one initial point (x of image ₀, y ₀) set out and line by line each pixel is scanned, in the time running into p (x, y) and be not 1, explanation is image background or has identified pixel and skip this pixel; In the time that to run into p (x, y) be 1 pixel, be counter C value by this pixel logo, and the adjacent all pixel values of the scanning element pixel that is 1 its value is designated to C, C value added to 1 after completing mark, continue next pixel logo.Identify pixel one by one until all pixels are scanned complete position, finally obtain one side corn kernel number.

Beneficial effect of the present invention comprises: (1) utilizes computer graphics disposal technology to replace manual corn ear seed counting method, can draw quick, easy, exactly corn ear seed sum, overcome in prior art scheme corn ear seed gate time long, subjectivity is strong, the large deficiency that waits of reading out data error; (2) the present invention is by using the part one side corn ear seed image collecting as sample, do regretional analysis according to one side corn ear kernal number in image and actual corn ear seed sum, obtain corn ear seed appraising model y=ax+b, and according to Corn Seeds appraising model, can estimate corn ear seed sum, rate of accuracy reached to 93% by one side corn kernel number in image; (3) the present invention adopts many threshold segmentation methods that RBGM and OTSU combine for processing in the many threshold segmentation methods of traditional OSTU along with segmentation threshold number is increased to many threshold values from single threshold, the calculated amount that optimal threshold is solved will become the problem of geometric growth, adopt the method can under same precision, greatly reduce separation calculation amount, and then shorten the separation calculation time.

Embodiment bis-

Based on identical design, the present invention also provides a kind of corn ear seed counting assembly, and as shown in Figure 5, this device comprises:

MBM 501, for obtaining corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Image pretreatment module 502, for obtaining the one-sided image of corn, and carries out image pre-service to described one-sided image, obtains corresponding gray level image;

Image is cut apart module 503, carries out many Threshold segmentations for the gray level image that image pretreatment module 502 is obtained, and obtains corresponding bianry image;

Counting module 504, for counting, obtain the number with the connected region that the gray-scale value for representing iblet is corresponding to the connected region with same grayscale value;

Computing module 505, the iblet appraising model obtaining according to the number getting in counting module 504 and MBM 501 is determined whole fruit ear kernal numbers.

Preferably, MBM 501 is specifically for obtaining following formula as corn ear seed appraising model:

y=ax+b；

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

Preferably, image pretreatment module 502 specifically comprises:

First processes submodule, for described one-sided image is carried out to medium filtering processing;

Second processes submodule, carries out Wallis filtering processing for the image that the first processing submodule is obtained;

The 3rd processes submodule, carries out histogram enhancement processing for the image that the second processing submodule is obtained, and obtains the gray level image that described one-sided image is corresponding.

Preferably, image is cut apart module 503, specifically for taking the gray level image that many threshold segmentation methods that RBGM and OTSU combine obtain image pretreatment module 502 to carry out the bianry image that corn kernel is cut apart and then obtained to image.

Preferably, cut apart module map and look like to cut apart module for taking gray level image that many threshold segmentation methods that RBGM and OTSU combine obtain image pretreatment module 502 to carry out image when cutting apart and then obtaining the bianry image of corn kernel, specifically for carrying out:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold

for making

get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\&eta;}}_m\frac{\&PartialD;{\mathrm{\&sigma;}}_B^2}{\&PartialD;t_i},i=\mathrm{1,2},...,N;$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\&sigma;}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&omega;}}_j{({\mathrm{\&mu;}}_k-{\mathrm{\&mu;}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart;

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, do not departing under the prerequisite of the technology of the present invention principle; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (10)

1. a corn ear seed method of counting, is characterized in that, comprising:

Step S1, obtains corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Step S2, obtains the one-sided image of corn, and described one-sided image is carried out to image pre-service, obtains corresponding gray level image;

Step S3, carries out many Threshold segmentations to the gray level image obtaining in step S2, obtains corresponding bianry image;

Step S4, counts the connected region with same grayscale value, obtains the number with the connected region that the gray-scale value for representing iblet is corresponding;

Step S5, determines whole fruit ear kernal numbers according to the iblet appraising model in the number getting in step S4 and step S1.

2. the method for claim 1, is characterized in that, described step S1 specifically comprises:

Obtain following formula as corn ear seed appraising model:

y=ax+b；

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

3. the method for claim 1, is characterized in that, described step S2 comprises:

Step S21, carries out medium filtering processing to described one-sided image;

Step S22, carries out Wallis filtering processing to the image obtaining in step S21;

Step S23, carries out histogram enhancement processing to the image obtaining in step S22, obtains the gray level image that described one-sided image is corresponding.

4. the method for claim 1, is characterized in that, described step S3, specifically comprises:

Take many threshold segmentation methods that RBGM and OTSU combine to carry out to the gray level image obtaining in step S2 the bianry image that corn kernel is cut apart and then obtained to image.

5. method as claimed in claim 4, is characterized in that, described in take many threshold segmentation methods that RBGM and OTSU combine to carry out to corn ear seed gray level image the bianry image that corn kernel is cut apart and then obtained to image, specifically comprise:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold for making get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\&eta;}}_m\frac{\&PartialD;{\mathrm{\&sigma;}}_B^2}{\&PartialD;t_i},i=\mathrm{1,2},...,N;$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\&sigma;}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&omega;}}_j{({\mathrm{\&mu;}}_k-{\mathrm{\&mu;}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart;

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

6. a corn ear seed counting assembly, is characterized in that, comprising:

MBM, for obtaining corn ear seed appraising model; Wherein said corn ear seed appraising model is for the funtcional relationship between the fruit ear kernal number that represents corn one side and whole fruit ear kernal numbers;

Image pretreatment module, for obtaining the one-sided image of corn, and carries out image pre-service to described one-sided image, obtains corresponding gray level image;

Image is cut apart module, carries out many Threshold segmentations for the gray level image that described image pretreatment module is obtained, and obtains corresponding bianry image;

Counting module, for counting, obtain the number with the connected region that the gray-scale value for representing iblet is corresponding to the connected region with same grayscale value;

Computing module, the iblet appraising model obtaining according to the number getting in described counting module and described MBM is determined whole fruit ear kernal numbers.

7. device as claimed in claim 6, is characterized in that, described MBM is specifically for obtaining following formula as corn ear seed appraising model:

y=ax+b；

Wherein, x represents the corn kernel number in the one-sided image of corn ear, and a, b are trained values, and y is whole fruit ear kernal numbers.

8. device as claimed in claim 6, is characterized in that, described image pretreatment module comprises:

First processes submodule, for described one-sided image is carried out to medium filtering processing;

Second processes submodule, carries out Wallis filtering processing for the image that the first processing submodule is obtained;

The 3rd processes submodule, carries out histogram enhancement processing for the image that the second processing submodule is obtained, and obtains the gray level image that described one-sided image is corresponding.

9. device as claimed in claim 6, it is characterized in that, described image is cut apart module, specifically for taking the gray level image that many threshold segmentation methods that RBGM and OTSU combine obtain described image pretreatment module to carry out the bianry image that corn kernel is cut apart and then obtained to image.

10. device as claimed in claim 9, it is characterized in that, described image is cut apart gray level image that module obtains described image pretreatment module for many threshold segmentation methods of taking RBGM and OTSU and combining and is carried out image when cutting apart and then obtaining the bianry image of corn kernel, specifically for carrying out:

Step S31, selects primary iteration point t _iwith iteration step length η _mfollowing many Threshold segmentations iterative formula is carried out to iterative, obtain optimal segmenting threshold

described optimal segmenting threshold

for making get peaked segmentation threshold:

$x_i^{\mathrm{mN}+j+1}=x_i^{\mathrm{mN}+j}-{\mathrm{\&eta;}}_m\frac{\&PartialD;{\mathrm{\&sigma;}}_B^2}{\&PartialD;t_i},i=\mathrm{1,2},...,N;$

Wherein,

j＝0,1,...,N-1;m＝0,1,...

${\mathrm{\&sigma;}}_B^2=\underset{k=0}{\overset{n-1}{\mathrm{\&Sigma;}}}\underset{j=k+1}{\overset{n-1}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_k{\mathrm{\&omega;}}_j{({\mathrm{\&mu;}}_k-{\mathrm{\&mu;}}_j)}^2;$

Wherein ω _k, μ _kbe respectively corresponding probability and the inequality of k class problem while representing n class problem to cut apart;

Step S32, uses the optimal segmenting threshold obtaining in step S31

the gray level image obtaining in step S2 is carried out to many Threshold segmentations.

Images