Plant gus signal recognition algorithm based on image processing technology

Abstract

The application discloses a plant GUS signal recognition algorithm based on an image processing technology, extracts image contour information and color difference information in a mode of image processing, obtains a plant main body and a GUS signal expression area in an image, and finally judges whether GUS signal expression occurs or not. The algorithm has low system resource occupation, is loose in the requirement for the quality of the plant image, and is accurate in the recognition of the GUS signal expression area. The algorithm overcomes the problems of unclear plant contour in a plant GUS staining picture and the interference of GUS positive signal area detection in the prior art.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of biological image processing, and particularly relates to a plant GUS signal recognition algorithm based on image processing technology. BACKGROUND

[0002] The GUS report system is one of the most commonly used tools in the field of plant molecular biology, which refers to using the uidA gene (GUS) derived from Escherichia coli as a reporter gene to detect the function and activity of a target gene or a regulatory element, and has a history of more than 30 years. The general usage is to connect the sequence with the target plant gene or regulatory element sequence to be detected to construct a vector, and then introduce the vector into a plant body, and the activity of the corresponding gene and regulatory element can be determined by detecting the amount and activity of the expressed GUS enzyme. At present, the GUS report system is still an important means for studying gene expression and function, promoter activity, protein stability and the like in plants due to its easy operation, high sensitivity and stable results.

[0003] The histochemical staining results of the GUS expression plant are mainly analyzed qualitatively by naked eye observation. Quantitative analysis can be performed by a fluorescence spectrophotometer and the like. The following defects exist:

[0004] The qualitative analysis method such as naked eye observation has the defects of subjective error, poor repeatability and inconvenience for comparison.

[0005] The fluorescence spectrophotometer method has the defects of complicated operation, high cost and poor repeatability of results.

[0006] The above existing methods also have the problem of universality, that is, the analysis process of the image of a certain plant is difficult to be directly popularized on GUS staining samples of other plants. SUMMARY

[0007] In view of the problems in the background art, in order to overcome the problems of unclear plant contour of the plant GUS staining picture and easy interference of the GUS positive signal area detection, the present application proposes a plant GUS signal recognition algorithm based on image processing technology, which can accurately recognize the presence or absence, intensity and distribution of the GUS signal in the plant image.

[0008] TECHNICAL SCHEME

[0009] A plant GUS signal recognition algorithm based on image processing technology, which comprises the following steps:

[0010] S1, acquiring an original image, including a reference image and a plant image to be detected, wherein the reference image includes a plant image without beta-glucuronidase (GUS) signal expression as a negative control and a plant image with GUS signal expression as a positive control;

[0011] S2, the original image is subjected to white balance correction to obtain standard image data, and a mark is added to the background and the plant main body part in the standard image data respectively;

[0012] S3, for the standard image data, a closed-form matting operation is performed according to the marks of the background and the plant main body part of the image to obtain an alpha matrix;

[0013] S4, the alpha matrix is binarized by using an OSTU operation to obtain a binarization matrix as a mask matrix main for marking the plant main body part;

[0014] S5, the plant main body part in the standard image data is compared with the plant main body part in the reference image in the standard image data in terms of blue channel pixel value to obtain a reference-corrected mask matrix main;

[0015] S6, the standard image data is split into red, green and blue channels, and then the values of different channels of the same pixel are compared, and when the red channel value and the green channel value are both greater than the blue channel value, the pixel is marked as "yellow" to obtain a yellow mask matrix yellow;

[0016] S7, the yellow mask matrix yellow is subjected to a logical and operation with the reference-corrected mask matrix main to obtain a mask matrix result which truly represents a GUS signal area;

[0017] S8, the blue channel of the plant image to be detected in the standard image data is inverted and subjected to a logical and operation with the mask matrix result to extract a truly expressed part, and subjected to a logical and operation with the mask matrix main to obtain a plant main body part, the number of pixels not equal to 0 in the two parts is counted respectively to obtain an expression area of the plant GUS signal, the average value of the values of the pixels not equal to 0 in the two parts is calculated respectively to obtain an expression amount, and the expression proportion is obtained by using the area of the truly expressed part divided by the area of the plant main body part.

[0018] Preferably, the reference image and the plant image to be detected are subjected to white balance correction before use to obtain standard image data.

[0019] Specifically, in S2, a black circular mark with a diameter of 1% of the image width is used to mark the inside of the plant main body in the image, and a white rectangular mark with a line width of 1% of the image width is used to mark the outside of the plant main body in the image.

[0020] Specifically, in S3, according to the marked image, the alpha matrix of the relationship between the plant main part and the background of the image is obtained by using the closed-form matting operation, wherein 1 indicates that the pixel point at the image position corresponding to the value belongs to the plant main part, 0 indicates that it belongs to the background part, and 0 and 1 indicate that the pixel point is a mixture of the plant main part and the background.

[0021] Specifically, in S4, the alpha matrix is converted into an integer type matrix, and then the OTSU method is used for binaryzation processing, wherein greater than or equal to 1 indicates the plant main part, and equal to 0 indicates the background part, to obtain the mask matrix main.

[0022] Specifically, the specific steps in S5 are as follows: comparing the relationship between the plant main part of the to-be-detected plant image and the plant main part blue channel pixel value in the reference, when the blue channel pixel value of the main part of the to-be-detected plant image is greater than the average value of the blue channel pixel of the negative control in the reference, the value of the pixel corresponding to the main matrix is reset to 0, so as to obtain the corrected mask matrix main.

[0023] Specifically, the acquisition steps of the yellow mask matrix yellow in S6 are as follows:

[0024] S6-1, a matrix yellow all of which is 0 is generated and is consistent with the size of the original image;

[0025] S6-2, the original image is split into red, green and blue channels;

[0026] S6-3, the values corresponding to the three channels of each pixel are compared respectively; when the red channel value and the green channel value of a certain pixel are both greater than the blue channel value, the value of the pixel corresponding position in the yellow matrix is set to 1.

[0027] Specifically, in S8, the blue channel of the original image of the to-be-detected plant image is inverted, which is realized by subtracting 255 from the blue channel matrix.

[0028] Advantages of the present application

[0029] The present application extracts image contour information and color difference information by image processing, obtains the plant main part and the GUS signal expression region in the image, and finally judges whether the GUS signal expression occurs or not. The algorithm system occupies low system resources, has a loose requirement on the quality of the plant image, and accurately identifies the GUS signal expression region.

[0030] The present application creatively uses the yellow mask matrix to modify the plant blue part in the original image, so as to obtain the plant GUS signal identification data, and overcomes the problems of unclear plant contour in the background technology and easy interference of the GUS positive signal region detection. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 Flow chart of the method of the present application DETAILED DESCRIPTION

[0032] The present application is further illustrated by the following examples, but the scope of the present application is not limited thereto:

[0033] In combination Figure 1 A plant GUS signal recognition algorithm based on image processing technology, which comprises the following steps:

[0034] S1, obtaining original images, including reference images and plant images to be detected, wherein the reference images include plant images without β-glucuronidase (GUS) signal expression as negative control and plant images with GUS signal expression as positive control;

[0035] In a preferred embodiment, before using the reference images and the plant images to be detected, white balance correction is performed to obtain standard image data.

[0036] S2, the original images are subjected to white balance correction to obtain standard image data, and labels are added to the background and the main part of the plant in the standard image data, respectively;

[0037] In a preferred embodiment, a 1% image width diameter black circular label is used to label the inside of the plant main part in the image, and a 1% image width line width white rectangular label is used to label the outside of the plant main part in the image.

[0038] S3, for the standard image data, according to the labels of the background and the main part of the plant, closed-form matting operation is performed to obtain an alpha matrix;

[0039] In a preferred embodiment, according to the labeled image, closed-form matting operation is used to obtain an alpha matrix of the relationship between the plant main part and the background in the image, wherein 1 in the matrix indicates that the pixel point at the image position corresponding to the value belongs to the plant main part, 0 indicates that it belongs to the background part, and 0 and 1 indicate that the pixel point is a mixture of the plant main part and the background.

[0040] S4, the alpha matrix is binarized using OSTU operation to obtain a binarization matrix as a mask matrix main of the labeled plant main part;

[0041] In a preferred embodiment, the alpha matrix is converted into an integer type matrix, and then OTSU method is used for binarization processing, wherein greater than or equal to 1 indicates the plant main part, and equal to 0 indicates the background part, to obtain a mask matrix main.

[0042] S5. Compare and correct the relationship between the blue channel pixel values ​​of the main plant part in the standard image data and the main plant part in the reference image in the standard image data to obtain the reference corrected mask matrix main.

[0043] The specific steps are as follows: compare the relationship between the blue channel pixel values ​​of the main body of the plant image to be detected and the blue channel pixel values ​​of the main body of the plant in the reference. When the blue channel pixel value of the main body of the plant image to be detected is greater than the average blue channel pixel value of the negative control in the reference, reset the value of the corresponding pixel in the main matrix to 0; thereby obtaining the corrected mask matrix main.

[0044] S6. Split the standard image data into three channels: red, green, and blue. Then compare the values ​​of the same pixel in different channels. When the values ​​of the red channel and the green channel are both greater than the values ​​of the blue channel, mark it as "yellow" to obtain the yellow mask matrix.

[0045] In a preferred embodiment, the steps for obtaining the yellow mask matrix are as follows:

[0046] S6-1. Generate a matrix yellow containing all zeros with the same size as the original image;

[0047] S6-2. The original image is split into three channels: red, green, and blue.

[0048] S6-3. Compare the values ​​of the three channels for each pixel; when the red channel value and green channel value of a pixel are both greater than the blue channel value, set the value of the corresponding position in the yellow matrix to 1.

[0049] S7. Perform a logical AND operation between the yellow mask matrix yellow and the mask matrix main after reference correction to obtain the mask matrix result that truly represents the GUS signal region.

[0050] S8. After inverting the blue channel of the plant image to be detected in the standard image data (the inversion step is achieved by subtracting the blue channel matrix from 255), perform a logical AND operation with the mask matrix result to extract the true expression part; perform a logical AND operation with the mask matrix main to obtain the main part of the plant; count the number of non-zero pixels in both parts to obtain the expression area of ​​the plant GUS signal; calculate the average value of the non-zero pixel values ​​in both parts to obtain the expression level; divide the area of ​​the true expression part by the area of ​​the main part of the plant to obtain the expression ratio.

[0051] Explanation of relevant technical terms:

[0052] GUS: β-glucuronidase.

[0053] GUS reporter system: A molecular biology tool to detect the function and activity of a gene or regulatory element of interest using the uidA gene (GUS) derived from E. coli as a reporter gene.

[0054] GUS plant image: An image of a plant expressing a GUS reporter system that has been stained is captured using an image capture device.

[0055] Mask: A matrix containing 0s or Is used to mark whether a pixel in a target image is applied to a calculation.

[0056] The specific embodiments described herein are merely illustrative of the spirit of the application. Various modifications or additions can be made to the specific embodiments described herein by those skilled in the art without departing from the spirit of the application or the scope of the appended claims.

Claims

1. A plant GUS signal recognition method based on image processing technology, characterized by It comprises the following steps: S1, obtaining original images, including reference images and plant images to be detected, wherein the reference images include plant images without beta-glucuronidase GUS signal expression as negative controls, and plant images with GUS signal expression as positive controls; S2, white balance correction is performed on the original images to obtain standard image data, and labels are added to the background and plant main part in the standard image data respectively; S3, for the standard image data, closed-form matting operation is performed according to the labels of the background and plant main part of the image to obtain an alpha matrix; S4, the alpha matrix is binarized using the OSTU operation to obtain a binarization matrix, which is used as a mask matrix main for marking the plant main part; S5, the plant main part in the standard image data is compared and corrected with the plant main part in the reference image in the standard image data in terms of blue channel pixel value to obtain a reference-corrected mask matrix main; S6, the standard image data is split into red, green and blue channels, and then the values of the same pixel in different channels are compared, and when the red channel value and the green channel value are both greater than the blue channel value, the pixel is marked as "yellow", and a yellow mask matrix yellow is obtained; S7, the yellow mask matrix yellow is subjected to logical and operation with the reference-corrected mask matrix main to obtain a mask matrix result which truly expresses the GUS signal area; S8, the blue channel of the plant image to be detected in the standard image data is inverted and subjected to logical and operation with the mask matrix result to extract the truly expressed part, and subjected to logical and operation with the mask matrix main to obtain the plant main part, the number of pixels not equal to 0 in the two parts is counted respectively to obtain the expression area of the plant GUS signal, and the average value of the values of the pixels not equal to 0 in the two parts is calculated respectively to obtain the expression amount, and the expression ratio is obtained by dividing the area of the truly expressed part by the area of the plant main part.

2. The method of claim 1, wherein Before using the reference images and the plant images to be detected, white balance correction is performed to obtain standard image data.

3. The method of claim 1, wherein In S2, a black circular mark with a diameter of 1% of the image width is used to mark the inside of the plant main part in the image, and a white rectangular mark with a line width of 1% of the image width is used to mark the outside of the plant main part in the image.

4. The method of claim 1, wherein In S3, according to the marked image, closed-form matting operation is used to obtain an alpha matrix of the relationship between the plant main part and the background of the image, wherein 1 in the matrix indicates that the pixel point corresponding to the image position belongs to the plant main part, 0 indicates that it belongs to the background part, and 0 and 1 indicate that the pixel point is a mixture of the plant main part and the background.

5. The method of claim 1, wherein In S4, the alpha matrix is converted into an integer type matrix, and then the OTSU method is used for binarization processing, wherein greater than or equal to 1 indicates the plant main part, and equal to 0 indicates the background part, to obtain a mask matrix main.

6. The method of claim 1, wherein The specific step in S5 is: comparing the plant main part of the plant image to be detected with the relationship of the blue channel pixel value of the plant main part in the reference, when the blue channel pixel value of the plant main part of the plant image to be detected is greater than the average value of the blue channel pixel of the negative control in the reference, resetting the value of the corresponding pixel of the main matrix to 0; thereby obtaining the corrected mask matrix main.

7. The method of claim 1, wherein The obtaining step of the yellow mask matrix yellow in S6 is: S6-1, generating a matrix yellow all of which is 0 and consistent with the size of the original image; S6-2, splitting the original image into red, green and blue channels; S6-3, comparing the values of the three channels corresponding to each pixel respectively; When the red channel value and the green channel value of a certain pixel are greater than the blue channel value at the same time, the value of the pixel corresponding position in the yellow matrix is set to 1.

8. The method of claim 1, wherein In S8, the original image blue channel of the plant image to be detected is inverted, which is realized by subtracting the blue channel matrix by 255.

Images