CN112668404B - Effective identification method for soybean diseases and insect pests - Google Patents

Abstract

The invention discloses an effective identification method for soybean diseases and insect pests, which solves the problems of hysteresis and inaccuracy in the aspects of soybean disease and pest identification and identification in the prior art based on field visual observation, laboratory analysis and the like. Establishing a database by utilizing known soybean plant characteristic maps under the influence of different types of soybean diseases and insect pests, extracting and storing various identification parameters related to the soybean plants under the influence of each disease and pest in the database at different growth periods, endowing different identification characteristic parameters with different weights according to the difference of the types of the diseases and the difference of the influences on the soybean plants, and establishing a distinguishing model of the types of the diseases and pests according to the extraction result of the characteristic parameters; the method and the device have the advantages of real-time performance and quasi-real-time performance.

Description

An effective identification method for soybean diseases and insect pests

technical field

The invention relates to an effective identification method for soybean diseases and insect pests.

Background technique

Soybean, as my country's main food crop and cash crop, plays a very important role in national economic production, especially in agricultural production. Soybean diseases and insect pests are not only one of the most important agricultural disasters in my country, but also in the world. Up to now, there have been more than 120 kinds of soybean diseases and insect pests reported in the world, including fungal diseases, bacterial diseases, virus diseases, insect diseases, etc. Different types of diseases and parasitic plant diseases, different soybean diseases and insect pests have different ways, types and impacts, but different diseases and insect pests will seriously affect the growth of soybeans, resulting in a significant reduction in soybean yield or even extinction. High-level and wide-ranging pests and diseases have caused huge losses to the national economy of my country and even all countries in the world, especially the national economy of countries that mainly rely on soybean production and planting. The disease and insect pests of soybeans are reflected in the average yield of crops suffering from pests and diseases by 1 to 20%, and even in severe cases, the production will cease to exist; after suffering from diseases and insect pests, the taste and quality of crops will be reduced to varying degrees, and the starch and protein content will also be greatly reduced; at the same time Soybeans affected by certain diseases and insect pests will cause poisoning, disease and even death of the people and animals they eat; in addition, because the diseases and insect pests usually have a large spread, they are often in the same area, and due to the influence of the disease and insect pests, the area cannot continue to be carried out. In addition, pest-affected soybeans are often difficult to store and transport, more perishable than healthy soybeans, and varieties with dangerous pests and diseases are more susceptible to export restrictions.

The effects of diseases and insect pests on soybean are mainly reflected in the rhizomes, leaves and fruits. For the roots of soybeans, the presence of pests and diseases will cause hairy roots, black roots, and rotten roots in soybeans, and the number of root nodules will decrease, which will affect the absorption of water, organic matter and other nutrients by the root system of crops, causing soybean plants to shrink and yellow. , slow growth; for soybean stems, the presence of pests and diseases will cause local lesions such as anthracnose and ulcers on the stems, which will affect the transportation of water and nutrients by soybean stems, resulting in soybean plants withering, rot, lodging and wilting. For soybean leaves, the presence of diseases and insect pests will cause discoloration and fading phenomena such as chlorosis, yellowing, and purplish discoloration, as well as disease spots such as round spots, angular spots, and stripes, thereby reducing soybean leaves. The photosynthetic efficiency greatly reduces the yield and quality of crops; for soybean fruits, the presence of diseases and insect pests will lead to the appearance of necrotic spots on the surface of the pods, strange and deformed soybean seeds, soft rot and internal dry rot of soybean seeds, etc. It directly affects the quality of soybeans, reduces its taste, and reduces the yield of soybeans.

Therefore, the identification of diseases and insect pests and the efficient and accurate identification of the types of diseases and insect pests are of great importance for determining the pathogen, disease course, yield assessment and economic loss caused by diseases and insect pests, as well as follow-up control measures for the types of diseases and insect pests. practical significance and guiding value. However, the existing methods for identification and identification of soybean diseases and insect pests are often based on the Koch-type rule through visual field observation and preliminary diagnosis of symptoms on the infected plant parts, combined with laboratory microscopic measurements and laboratory microscopic measurements of infected plants. Analysis of physical and chemical parameters to determine the types and causes of pests and diseases, this kind of measurement often requires professional practitioners, and the discrimination process has a certain lag, and there may be misdiagnosis or missed diagnosis in the measurement process, thus delaying the best results. During the treatment period, greater losses are caused to soybeans affected by pests and diseases.

SUMMARY OF THE INVENTION

Based on the above deficiencies, the present invention provides a new effective identification method for soybean diseases and insect pests, which solves the hysteresis and inconsistency in the existing identification and identification of soybean diseases and insect pests in the prior art based on field visual observation and laboratory analysis. exact question.

The technology adopted in the present invention is as follows: an effective identification method for soybean diseases and insect pests, the steps are as follows:

Step 1. Standardized acquisition of soybean organ images

Acquiring standardized images of soybeans affected by diseases and insect pests in different growth periods involves the acquisition of standardized images of leaf, root and grain parts of soybean plants affected by diseases and insect pests;

Step 2. Standardization of soybean organ images

Geometric distortion correction is carried out on the standardized photos obtained from soybean organ images in different growth periods to remove the geometric distortions brought by the photographing process, and a uniform size is selected for different organ images for cropping operations, including standardized images of soybean plants, leaves, Normalized image of root system, normalized image of grain;

Step 3. Acquisition of soybean disease and insect pest identification parameters

This step involves the extraction of characteristic parameters of soybean organs affected by diseases and insect pests in different growth periods, including the acquisition of characteristic parameters of soybean leaves, the acquisition of characteristic parameters of soybean roots, and the acquisition of characteristic parameters of soybean grains;

Step 4. Establishment of database of soybean pests and diseases

A database is established using the known characteristic maps of soybean plants under the influence of different types of soybean diseases and insect pests. In the database, the various characteristic parameters involved in step 3 are extracted and saved for each type of soybean plant under the influence of diseases and insect pests under different growth periods. , assign different weights to different characteristic parameters according to the difference of the types of diseases and insect pests and their effects on plants, and establish the discriminant model of the types of diseases and insect pests according to the extraction results of the characteristic parameters;

Step 5. Identification and identification of soybean diseases and insect pests

Using the characteristic parameters of plants affected by different types of soybean diseases and insect pests in the database under different growth periods and the discriminant model of the types of diseases and insect pests in the database to identify and identify the types of soybean diseases and insect pests that are actually affected, extract the characteristic parameters of soybean plants to be identified. The results are compared and calculated with the characteristic parameters of the plants under each disease and insect pest existing in the database. Finally, the calculation results of the discriminant model of the characteristic parameters of the soybean plants to be identified are compared with the calculation results of the discriminant model of the plants under each disease and insect pest to compare, calculate and match. So as to realize the identification and identification of soybean diseases and insect pests.

The present invention also has the following technical features:

1. Step 3 as described above: further process the standardized images of leaves obtained during the four key growth stages of soybean plants affected by diseases and insect pests in the flowering stage, pod setting stage, drumming stage and maturity stage. Corresponding characteristic parameters were extracted from the standardized images of soybean leaves that identified the types of diseases and insect pests. For the leaves of soybean plants affected by the diseases and insect pests to be identified in a specific growth period, in order to remove the heterogeneity of the extraction results of characteristic parameters among samples, 20 plants were selected, each Each plant randomly selects a standardized color image of a leaf, extracts the image matrices R(x,y), G(x,y), B(x,y) of the red, green, and blue components of the standardized color image of the leaf, and then Convert the normalized color image of the leaf to a grayscale image according to the formula F(x,y)=R(x,y)/3+G(x,y)/3+B(x,y)/3 for the three components , perform histogram statistics on the grayscale map, find the grayscale value t1 corresponding to the lowest frequency point between the two grayscale peaks of the green part of soybean leaves and the discolored part affected by diseases and insect pests, and set all pixels whose grayscale is greater than the threshold t1. The brightness value is set to 0, and the brightness value of many pixels whose grayscale is less than the threshold t1 is set to 1, so as to convert the grayscale image into a binary image; The measurement result w1' cm, calculate the actual side length of each pixel z1=w1'/w1 cm, the actual area of each pixel is z1 ² square cm, according to the actual side length of the pixel z1 cm, the size of the pixel The actual area z1 is ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of each spot, calculate the average value of the leaf characteristic parameters of the plant affected by the disease and insect pest to be identified in this specific growth period, including the ratio of the leaf spot area of the plant to be identified, a1, The number of spots on the leaves of the plants to be identified is a2, the area value of the spots on the leaves of the plants to be identified is a3, and the perimeter value of the spots on the leaves of the plants to be identified is a4; once again, perform histogram statistics on the grayscale images of the plants affected by diseases and insect pests to further extract those affected by diseases and insect pests The overall morphological characteristics of leaves, find the gray value t1' corresponding to the lowest gray level frequency between the two gray peaks of soybean leaf and background color, and set the brightness value of all pixels whose gray value is greater than the threshold t1' to 0 , set the brightness value of all pixels whose gray level is less than t1' to 1, realize the binarization processing of the whole soybean leaf, and extract the number of pixels corresponding to the long axis and the number of pixels corresponding to the short axis of the leaf according to the leaf binary map Calculate the leaf length-to-width ratio a5 of the specific growth expectation identification plant with pests and diseases and the leaf ellipticity a6 of the specific growth expectation identification plant with pests and diseases;

For the roots of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters between samples, 20 standardized root images of plants were selected, and the red, green and blue components of the leaf color image were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 Convert the color leaf image to grayscale, perform histogram statistics on the root grayscale image, and find the grayscale between the two grayscale peaks of soybean root and background color The grayscale value t2 corresponding to the lowest frequency point, set the brightness value of all pixels whose grayscale is greater than the threshold t2 to 0, and set the brightness value of all pixels whose grayscale is less than t2 to 1, so as to convert the root grayscale image into two Value map, according to the pixel value w2 corresponding to the width of the root image and the measurement result of the plant corresponding to the root image w2' cm, calculate the actual side length of each pixel z2=w2'/w2 cm, the actual area of each pixel Then it is z2 ² square centimeters. According to the actual side length of the pixel z2 cm, the actual area of the pixel z2 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of the root system of plants affected by pests and diseases, the identification of the specific growth expectation is calculated. The root area value of the plant affected by diseases and insect pests is b1, and the number of root nodules of the plant to be identified is b2. Calculate the average value of the root texture characteristic parameters of the plant affected by diseases and insect pests to be identified in the specific growth period according to the grayscale map. The specific calculation process of the root texture characteristic parameters is as follows: Calculate the root contrast texture feature value b3 of the plant with plant diseases and insect pests to be identified according to formula 1, the texture feature value b4 of the root system consistency of the plants to be identified according to formula 2, the texture feature value b5 of the root system entropy value of the plants to be identified according to formula 3, and the texture feature value b5 of the plant root system to be identified according to formula 4 The eigenvalue b6 of root energy texture of plants with pests and diseases,

Among them, p(i,j) represents the value of the grayscale co-occurrence matrix calculated from the root grayscale image at the i-th row and j-th column position, and represents the pixel and gray level of the gray level i in the root grayscale image. The probability that the pixels with level j appear at the same time in a fixed direction and pixel interval, n is the difference between the two gray levels of i and j in the root gray image, and Ng represents the gray co-occurrence matrix extracted from the root gray image series;

For the grains of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters among samples, 20 plants were selected, and the red, green and blue components of the color grain image of each plant's grain were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 Convert the color grain image into a grayscale image, and perform histogram statistics on the grayscale image to find the lowest frequency between the two grayscale peaks in the soybean kernel part and the background part of the test bench The gray value t3 corresponding to the point, the brightness value of all pixels whose gray value is greater than the threshold value t3 is set to 1, and the brightness value of many pixels whose gray value is less than the threshold value t3 is set to 0, so as to convert the grayscale image of the grain into two values. Figure, calculate the actual side length of each pixel z3=w3'/w3 cm according to the pixel value w3 corresponding to the width of the grain image affected by the disease and insect pests and the measurement result of the root image corresponding to the plant w3' cm, each pixel The actual area is z3 ² square centimeters, according to the actual side length of the pixel z3 centimeters, the actual area of the pixel z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by pests and diseases, calculate the specific The average value of the grain characteristic parameters of the plants affected by diseases and insect pests to be identified during the growth period, including the number of grains of the plants to be identified c1, the value of the grain area of the plants to be identified c2, the value of the grain circumference of the plants to be identified c3, and the length of the grains of the plants to be identified. The width ratio is c4, the grain curvature value of the plants to be identified with pests and diseases is c5, and the grain ellipticity value of the plants to be identified with pests and diseases is c6.

2. Step 4 as described above: use the images of soybean plants affected by all known types of diseases and insect pests in different growth stages of flowering, pod setting, drumming and maturity to establish a characteristic parameter database, and for a specific growth period. The extraction and database construction of soybean leaf characteristic parameters under the influence of known diseases and insect pests during the period were combined with the method of step 3 to convert the color leaf image under the influence of known diseases and insect pests under the specific growth period into a grayscale image, and the method of step 3 was used to convert the gray scale image. The degree map is used for histogram statistics, and the gray value T1 corresponding to the lowest frequency point between the two gray value peaks of the green part of the leaf and the discolored part affected by diseases and insect pests is found, and the brightness value of all pixels whose gray value is greater than the threshold T1 is set as 0, set the brightness value of many pixels whose grayscale is less than the threshold T1 to 1, to convert the grayscale image into a binary image, according to the pixel value W1 corresponding to the width of the leaf image and the measurement result W1 of the plant corresponding to the leaf image 'cm, calculate the actual side length of each pixel Z1=W1'/W1 cm, the actual area of each pixel is Z1 ² square cm, according to the actual side length of the pixel Z1 cm, the actual area of the pixel Z1 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters for each spot, calculate the average value of the plant leaf characteristic parameters of the known disease and insect pest types in this specific growth period, including the known disease and insect pest plant leaf spot area ratio A1, known disease and insect pests The number of leaf spots of plants is A2, the area value of leaf spots of plants with known diseases and insect pests is A3, and the perimeter value of leaf spots of plants with known diseases and insect pests is A4. The grayscale map of plants affected by pests and diseases is used for histogram statistics, and the grayscale value T1' corresponding to the lowest point of grayscale frequency between the two grayscale peaks of soybean leaf and background color is found, and all pixels whose grayscale is greater than the threshold T1' The brightness value is set to 0, and the brightness value of all pixels whose grayscale is less than T1' is set to 1, so as to realize the binarization processing of the whole soybean leaf, and extract the number of pixels corresponding to the long axis and the short axis of the leaf according to the leaf binary map. The corresponding pixel number is used to calculate the leaf spot aspect ratio value A5 of the plant with known plant diseases and insect pests and the ellipticity value of the plant spot with known plant diseases and insect pests in the specific growth period A6;

The extraction and database establishment of soybean root characteristic parameters under the influence of known diseases and insect pests in a specific growth period are combined with the method of step 3 to convert the color root system images under the influence of known diseases and insect pests in the specific growth period into grayscale images, and the steps The method of 3 performs histogram statistics on the grayscale image, and finds the grayscale value T2 corresponding to the lowest point of the grayscale frequency between the peaks of the two grayscale parts of the soybean root system and the background color. The value is set to 0, the brightness value of all pixels whose grayscale is less than T2 is set to 1, and the root grayscale image is converted into a binary image. According to the pixel value W2 corresponding to the width of the root image and the measurement of the plant corresponding to the root image Result W2' cm, calculate the actual side length of each pixel Z2=W2'/W2 cm, the actual area of each pixel is Z2 ² square cm, according to the actual side length of the pixel Z2 cm, the actual size of the pixel The area Z2 is ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant roots affected by plant diseases and insect pests. Calculate the mean value of root characteristic parameters of all plants affected by known plant diseases and insect pests, including known plant root system area value B1, known plant diseases and insect pests The number of root nodules in the plant root system is B2. According to the formula 1 in step 3, the root contrast texture characteristic value B3 of the known plant with plant diseases and insect pests is calculated, the root consistency texture characteristic value of the plant with known plant diseases and insect pests is calculated according to the formula 2, and the texture characteristic value of the known plant root system is calculated according to the formula 3. Root entropy value texture feature value B5, calculate root energy texture feature value B6 of known plant diseases and insect pests according to formula 4;

For the grains of soybean plants affected by known diseases and insect pests in a specific growth period, the color grain images under the influence of known diseases and insect pests in the specific growth period are converted into grayscale images in combination with the method in step 3, and the grayscale images are analyzed. The histogram statistics find the gray value T3 corresponding to the lowest frequency point between the two gray value peaks of the soybean grain part and the background part of the test bench, and set the brightness value of all pixels whose gray value is greater than the threshold T3 to 1, and the gray value is less than the threshold value. The brightness value of most pixels in T3 is set to 0, and the grayscale image of the grain is converted into a binary image. , calculate the actual side length of each pixel Z3=W3'/W3 cm, the actual area of each pixel is Z3 ² square cm, according to the actual side length of the pixel Z3 cm, the actual area of the pixel Z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by diseases and insect pests, calculate the average value of characteristic parameters of plant grains known to be affected by plant diseases and insect pests in this specific growth period, including the number of known plant grains C1, known The grain area value of plants with known diseases and insect pests is C2, the perimeter value of grains of plants with known diseases and insect pests is C3, the ratio of grain length to width of plants with known diseases and insect pests is C4, the curvature value of grains of plants with known diseases and insect pests is C5, and the ellipticity value of grains of plants with known diseases and insect pests is C6. To determine the difference in the effects _on soybean plant organs, determine the weight K ₁ of the leaf characteristic parameters of each known soybean pests and diseases, determine the weight K2 of the root characteristic parameters of each known soybean pests and diseases, and determine the weights of each known soybean pests and diseases. The weight K ₃ of the characteristic parameter of the type grain, and at the same time satisfy the normalized characteristic of the weight parameter, that is, K ₁ +K ₂ +K ₃ =1, according to the extraction results and Based on the extraction results of different organ image feature parameters of known pest species in the database, a pest and disease discrimination model is established. The discriminant model is shown in Equation 5:

4. Step 5 as described above: use the characteristic parameters of plants affected by different types of known soybean diseases and insect pests in the database in step 4 under different growth periods and the known disease and insect pest species discrimination models in the database to realize the identification of affected soybean diseases and insect pests. Identification and identification, extraction results a1, a2, a3, a4, a5, a6 of leaf characteristic parameters of a plant affected by diseases and insect pests to be identified; extraction results b1, b2, b3, b4, b5 of root characteristic parameters of plants affected by diseases and insect pests to be identified , b6; the extraction results c1, c2, c3, c4, c5, and c6 of the grain characteristic parameters of the plants affected by diseases and insect pests to be identified, the extraction results of the characteristic parameters of each soybean affected by diseases and insect pests to be identified and the The extraction results of the characteristic parameters of soybeans known to be affected by diseases and insect pests are brought into the discriminant model in step 4, and the P value of the model calculation results of the types of diseases and insect pests to be identified and all known types of diseases and insect pests is calculated according to the judgment model, and the smaller the calculation result of the P value is , which proves that the pest type to be identified is closer to the known pest and disease type corresponding to the calculation result of the P value, so as to realize the identification and identification of soybean pest and disease types.

The beneficial effects and advantages of the present invention: compared with the analysis and identification in the laboratory, the method involved in the present invention has less identification cost, low cost, less loss, less damage to plants, and has real-time and quasi-real-time performance. The identification personnel can quickly obtain the characteristic parameters of soybean plants to be identified with pests and diseases only through images, and realize the identification of pests and diseases according to the extraction results. In addition, the present invention has very low requirements for the professional knowledge of the identification personnel, and the identification personnel can use the method involved in the present invention only through simple training, and can realize the soybeans to be identified by utilizing a large amount of data of various known types of diseases and insect pests in the database. The types of pests and diseases are identified and identified, and the entire identification process does not require the identification personnel to have a comprehensive grasp of the knowledge of pests and diseases. In addition, compared with the traditional method for identifying the types of soybean diseases and insect pests through visual inspection in the field, the method involved in the present invention has a fast identification speed. Due to the high degree of automation in the extraction process of the plant parameters to be identified, the present invention can also batch process a large number of plants simultaneously by programming means. The soybean plant image of the type of diseases and insect pests to be identified is extracted, and the parameters of the soybean plant of the type of diseases and insect pests to be identified are extracted, so as to realize the rapid and efficient batch identification of the types of diseases and insect pests of soybean plants.

Description of drawings

Figure 1 is a grayscale image of leaves of plants affected by diseases and insect pests;

Fig. 2 is a binary map of the interior of leaves of plants affected by diseases and insect pests;

Figure 3. The overall binary image of leaves of plants affected by diseases and insect pests;

Fig. 4 Grayscale image of roots of plants affected by pests and diseases;

Fig. 5 Binary map of root system of plants affected by diseases and insect pests;

Figure 6 Grayscale image of grains of plants affected by diseases and insect pests;

Fig. 7 Binary image of grains of plants affected by diseases and insect pests;

Detailed ways

The present invention will be further described below according to the accompanying drawings of the description:

Example 1

Step 1: Obtain the standardized images of typical soybean plant organs affected by diseases and insect pests in different growth stages such as flowering stage, pod setting stage, drumming stage and maturity stage of soybean,

For images of soybean leaves affected by diseases and insect pests in a certain growth period, install a digital camera on a fixed bracket, place the leaves affected by diseases and insect pests on the test bench, adjust the position of the bracket so that the camera lens is perpendicular to the test bench, and the camera lens is far from the soybean leaves At a height of 20cm, use a photometer to determine and adjust the photographing environment, take pictures of the entire leaf affected by diseases and insect pests to obtain a standardized image of the leaf and take a checkerboard grid calibration plate image;

For soybean root images affected by diseases and insect pests in a certain growth period, install a digital camera on a fixed bracket, adjust the position of the camera so that the camera lens is perpendicular to the ground, lay the soybean plants on the ground, and ensure that the camera lens is at a height of 1m from the ground. The photometer determines and adjusts the photographing environment, and takes pictures of the large root system, which is used to obtain the standardized image of the soybean root system affected by diseases and insect pests and to take the image of the checkerboard grid calibration plate;

For an image of soybean grains affected by diseases and insect pests in a certain growth period, place the pods affected by diseases and insect pests on the test bench, adjust the position of the bracket so that the camera lens is perpendicular to the test bench, and then take out the soybean grains of the whole plant and place them on the test bench. , adjust the position of the bracket so that the camera lens is perpendicular to the test bench, and the camera lens is 20 cm from the height of the soybean grains, and the photometer is used to determine and adjust the photographing environment, take the standardized image of soybean particles affected by diseases and insect pests, and take the checkerboard grid calibration plate image.

Step 2: Standardize the image extraction results of various types of soybean organs affected by diseases and insect pests in step 1. For different types of soybean organ images, since the normalized images and the corresponding calibration plate images have the same shooting conditions, the standardized images of different types of soybean organs are The geometric distortion of the corresponding black-and-white checkerboard calibration plate image is also the same. In this embodiment, the black-and-white checkerboard grid calibration plate corresponding to each type of soybean organ adopts the polynomial correction method to correct the geometric distortion, and records the images of different organ types. Then, the standardized images of soybean organs under this type are geometrically corrected, and the appropriate size is selected to crop the soybean organ images under this type, so as to realize the standardized prediction of various types of soybean organ images affected by diseases and insect pests. Handling operations.

Step 3: Further process the standardized images of the leaves obtained in the four key soybean growth stages of the soybean plants affected by diseases and insect pests at the flowering stage, pod setting stage, drumming stage and maturity stage, and for each soybean leaf with the identified disease and insect pest type Normalize the image to extract the corresponding feature parameters.

For the leaves of soybean plants affected by the diseases and insect pests to be identified in a specific growth period, in order to remove the heterogeneity of the extraction results of characteristic parameters between samples, 20 plants were selected, and each plant randomly selected a standardized color image of a leaf. Extract the image matrices R(x,y), G(x,y), B(x,y) of the red, green, and blue components of the leaf standardized color image, and then press the formula F(x,y for the three components )=R(x,y)/3+G(x,y)/3+B(x,y)/3 Convert the normalized color image of the leaf into a grayscale image, and the grayscale image extraction result is shown in Figure 1 . Perform histogram statistics on the grayscale image to find the grayscale value t1 corresponding to the lowest frequency point between the two grayscale peaks of the green part of soybean leaves and the discolored part affected by diseases and insect pests, and set the brightness of all pixels whose grayscale is greater than the threshold t1. The value is set to 0, and the brightness value of many pixels whose grayscale is less than the threshold t1 is set to 1, so as to realize the conversion of the grayscale image into a binary image, as shown in Figure 2. According to the pixel value w1 corresponding to the width of the leaf image and the measurement result of the plant corresponding to the leaf image w1' cm, calculate the actual side length of each pixel z1=w1'/w1 cm, and the actual area of each pixel is z1 ² square centimeters, according to the actual side length of the pixel z1 cm, the actual area of the pixel z1 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of each spot, calculate the plant affected by the disease and insect to be identified in this specific growth period The average value of the leaf characteristic parameters, including the ratio a1 of the leaf spot area of the plant to be identified, the number of leaf spots of the plant to be identified a2, the value of the leaf spot area of the plant to be identified a3, and the value of the leaf spot perimeter of the plant to be identified a4.

In order to further extract the overall morphological characteristics of leaves affected by pests and diseases, the histogram statistics of the grayscale images of plants affected by pests and diseases were performed again, and the grayscale corresponding to the lowest point of grayscale frequency between the two grayscale peaks of soybean leaves and background color was found. value t1', set the brightness value of all pixels whose grayscale is greater than the threshold t1' to 0, and set the brightness value of all pixels whose grayscale is less than t1' to 1 to realize the binarization of the whole soybean leaf, as shown in the figure 3 is the result of binarization processing of the whole soybean leaf, extract the number of pixels corresponding to the long axis of the leaf and the number of pixels corresponding to the short axis of the leaf according to the binary map of the leaf to calculate the specific growth expectation and identify the plant leaf aspect ratio a5 of the disease and insect pest and the specific Growth expectation identification of plant leaf ellipticity a6 with pests and diseases.

For the roots of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters between samples, 20 standardized root images of plants were selected, and the red, green and blue components of the leaf color image were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 converts the color leaf image to grayscale, as shown in Figure 4.

Perform histogram statistics on the root gray image, find the gray value t2 corresponding to the lowest point of gray level frequency between the two gray peaks of soybean root and background color, and set the brightness value of all pixels whose gray scale is greater than the threshold t2 as If it is 0, set the brightness value of all pixels whose gray level is less than t2 to 1, so as to convert the root gray image into a binary image, as shown in Figure 5.

According to the pixel value w2 corresponding to the width of the root image and the measurement result of the plant corresponding to the root image w2' cm, calculate the actual side length of each pixel z2=w2'/w2 cm, and the actual area of each pixel is z2 ² square centimeters, according to the actual side length of the pixel z2 cm, the actual area of the pixel z2 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of the plant root system affected by the disease and insect pest, calculate the specific growth expectation to identify the root system of the plant with diseases and insect pests The area value b1, the number b2 of the root nodules of the plants to be identified with diseases and insect pests, the average value of the root texture feature parameters of the plants to be identified in the specific growth period is calculated according to the grayscale map, the specific calculation process of the root texture feature parameters is as follows, according to formula 1 Calculate the characteristic value b3 of the contrast texture of the root system of the plant to be identified, according to the formula 2, the texture characteristic value of the root system of the plant to be identified, b4, according to the formula 3, the entropy value of the root system of the plant to be identified, the texture characteristic value b5, and the root of the plant to be identified according to the formula 4 Energy texture feature value b6.

Among them, p(i,j) represents the value of the grayscale co-occurrence matrix calculated from the root grayscale image at the i-th row and j-th column position, and represents the pixel and gray level of the gray level i in the root grayscale image. The probability that the pixels with level j appear at the same time in a fixed direction and pixel interval, n is the difference between the two gray levels of i and j in the root gray image, and Ng represents the gray co-occurrence matrix extracted from the root gray image 's level.

For the grains of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters among samples, 20 plants were selected, and the red, green and blue components of the color grain image of each plant's grain were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 converts the color grain image into a grayscale image, as shown in Figure 6,

Perform histogram statistics on the grayscale image to find the grayscale value t3 corresponding to the lowest frequency point between the peaks of the two grayscale parts of the soybean grain part and the background part of the test bench, and set the brightness value of all pixels whose grayscale is greater than the threshold t3 to 1 , the brightness value of many pixels whose gray level is less than the threshold t3 is set to 0, and the grain gray level image is converted into a binary image, as shown in Figure 7.

Calculate the actual side length of each pixel z3=w3'/w3 cm according to the pixel number w3 corresponding to the width of the grain image affected by diseases and insect pests and the measurement result w3' cm of the corresponding plant in the root image, and the actual side length of each pixel is calculated. The area is z3 ² square centimeters. According to the actual side length of the pixel z3 cm, the actual area of the pixel z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by pests and diseases, the specific growth period is calculated. The average value of the grain characteristic parameters of the plants affected by diseases and insects to be identified, including the number of grains of the plants to be identified c1, the value of the grain area of the plants to be identified c2, the value of the grain circumference of the plants to be identified c3, and the ratio of the grains of the plants to be identified. c4, the grain curvature value c5 of the plants to be identified with pests and diseases, and the grain ellipticity value c6 of the plants to be identified with pests and diseases.

Step 4: Use the images of soybean plants affected by all known types of diseases and insect pests in different growth stages of flowering, pod setting, drumming and maturity to establish a feature parameter database, and the impact of known diseases and insect pests in a specific growth period. The extraction and database establishment of soybean leaf characteristic parameters are combined with the methods and formulas in step 3 to convert the color leaf image under the influence of known pests and diseases under the specific growth period into a grayscale image, and the method in step 3 is used to convert the grayscale image. Perform histogram statistics to find the gray value T1 corresponding to the lowest frequency point between the two gray value peaks of the green part of the leaf and the discolored part affected by diseases and insect pests, and set the brightness value of all pixels whose gray value is greater than the threshold T1 to 0, Set the brightness value of many pixels whose grayscale is less than the threshold T1 to 1, and convert the grayscale image into a binary image. , calculate the actual side length of each pixel Z1=W1'/W1 cm, the actual area of each pixel is Z1 ² square cm, according to the actual side length of the pixel Z1 cm, the actual area of the pixel Z1 ² square cm, and the number of pixels corresponding to the extraction results of different parameters of each spot, calculate the average value of the leaf characteristic parameters of plants with known pests and diseases in this specific growth period, including the ratio of leaf spot area of plants with known pests and diseases A1, and the leaves of plants with known pests and diseases. The number of spots is A2, the leaf spot area value of plants with known diseases and insect pests is A3, and the perimeter value of leaf spots of plants with known diseases and insect pests is A4. The histogram statistics of the plant grayscale map are used to find the grayscale value T1' corresponding to the lowest point of the grayscale frequency between the two grayscale peaks of soybean leaf and background color, and the brightness value of all pixels whose grayscale is greater than the threshold T1' is calculated Set to 0, set the brightness value of all pixels whose gray level is less than T1' to 1, realize the binarization processing of the whole soybean leaf, and extract the number of pixels corresponding to the long axis of the leaf and the corresponding short axis according to the leaf binary map. The number of pixels was used to calculate the leaf spot aspect ratio A5 of the plants with known plant diseases and insect pests and the ellipticity value A6 of the plant spots with known plant diseases and insect pests in the specific growth period.

Extraction and database establishment of soybean root characteristic parameters under the influence of known diseases and insect pests in a specific growth period Combined with the methods and formulas in step 3, the color root system images under the influence of known diseases and insect pests in the specific growth period are converted into grayscale images , use the method in step 3 to perform histogram statistics on the grayscale image, find the grayscale value T2 corresponding to the lowest point of the grayscale frequency between the two grayscale peaks of the soybean root system and the background color, and set the grayscale value greater than the threshold value T2. The brightness value of all pixels is set to 0, and the brightness value of all pixels whose grayscale is less than T2 is set to 1 to convert the root grayscale image into a binary image. According to the pixel value W2 and the root image corresponding to the width of the root image Corresponding to the measurement result of the plant W2' cm, calculate the actual side length of each pixel Z2 = W2'/W2 cm, the actual area of each pixel is Z2 ² square cm, according to the actual side length of the pixel Z2 cm, The actual area of the pixel is Z2 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of the root system of the plant affected by diseases and insect pests. Calculate the average value of the characteristic parameters of the root system of all plants affected by known diseases and insect pests, including the root area value of plants affected by known diseases and insect pests B1 , the number B2 of root nodules of plants with known plant diseases and insect pests is known, the characteristic value B3 of root contrast texture of plants with known plant diseases and insect pests is calculated according to the formula 1 in step 3, and the characteristic value B4 of the consistency texture of roots of plants with known plant diseases and insect pests is calculated according to formula 2 in step 3, Calculate the root entropy texture feature value B5 of plants with known plant diseases and insect pests according to formula 3 in step 3, and calculate root energy texture feature value B6 of plants with known plant diseases and insect pests according to formula 4 in step 3.

For the grains of soybean plants affected by known pests and diseases in a specific growth period, the color image of grains affected by known pests and diseases in the specific growth period is converted into a grayscale image by combining the method and formula in step 3. Figure 1. Perform histogram statistics to find the gray value T3 corresponding to the lowest frequency point between the two grayscale peaks of the soybean kernel part and the background part of the test bench. The brightness value of many pixels less than the threshold T3 is set to 0, and the grayscale image of the grain is converted into a binary image. 'cm, calculate the actual side length of each pixel Z3=W3'/W3 cm, the actual area of each pixel is Z3 ² square cm, according to the actual side length of the pixel Z3 cm, the actual area of the pixel Z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by diseases and insect pests, calculate the average value of the characteristic parameters of plant grains known to be affected by plant diseases and insect pests in this specific growth period, including the number of known plant grains C1, The grain area value of known plants with diseases and insect pests is C2, the value of grain perimeter of plants with known plant diseases and insect pests is C3, the value of grain aspect ratio of plants with known plant diseases and insect pests is C4, the value of grain curvature of plants with known plant diseases and insect pests is C5, and the ellipticity value of grains of plants with known plant diseases and insect pests is C6. According to the difference in the effects of different types of diseases and insect pests on soybean plant organs, determine the weight K ₁ of the leaf characteristic parameters of each known type of soybean diseases and insect pests, determine the weight K ₂ of the root characteristic parameters of each known type of soybean diseases and insect pests, and determine the The weight K ₃ of the grain characteristic parameters of the known soybean disease and insect pest types, and at the same time satisfies the normalized characteristics of the weight parameters, that is, K ₁ +K ₂ +K ₃ =1. According to the extraction results of different organ image feature parameters of the soybean pests and diseases to be identified and the extraction results of different organ image feature parameters of known pests and diseases in the database, a pest discrimination model is established, and the discrimination model is shown in Equation 5.

Step 5: Use the characteristic parameters of plants affected by different types of known soybean diseases and insect pests in the database in step 4 under different growth periods and the discrimination model of known types of soybean diseases and insect pests in the database to realize the identification and identification of the affected soybean diseases and insect pests. A1, a2, a3, a4, a5, a6 of leaf characteristic parameter extraction results of a plant affected by diseases and insect pests to be identified; extraction results b1, b2, b3, b4, b5, b6 of root system characteristic parameters of a plant affected by diseases and insect pests to be identified; The extraction results of grain characteristic parameters c1, c2, c3, c4, c5, c6 of the plants affected by diseases and insect pests. The extraction results of the characteristic parameters of each soybean affected by diseases and insect pests to be identified and the extraction results of characteristic parameters of each known soybean affected by diseases and insect pests under this growth period are brought into the discriminant model in step 4, and the to-be-determined model is calculated according to the judgment model. The size of the P value of the model calculation result of identifying the type of pest and disease and all known types of pests and diseases. The smaller the calculation result of the P value, the closer the type of pest to be identified is to the known type of pest and disease corresponding to the calculation result of the P value, so as to achieve Fast, accurate and effective identification and identification of soybean pest types.

Claims (3)

1. an effective identification method for soybean diseases and insect pests, is characterized in that, method steps are as follows: Step 1. Standardized acquisition of soybean organ images Acquiring standardized images of soybeans affected by diseases and insect pests in different growth periods involves the acquisition of standardized images of leaf, root and grain parts of soybean plants affected by diseases and insect pests; Step 2. Standardization of soybean organ images Geometric distortion correction is carried out on the standardized photos obtained from soybean organ images in different growth periods to remove the geometric distortions brought by the photographing process, and a uniform size is selected for different organ images for cropping operations, including standardized images of soybean plants, leaves, Normalized image of root system, normalized image of grain; Step 3. Acquisition of soybean disease and insect pest identification parameters This step involves the extraction of characteristic parameters of soybean organs affected by diseases and insect pests in different growth periods, including the acquisition of soybean leaf characteristic parameters, the acquisition of soybean root characteristic parameters, and the acquisition of soybean grain characteristic parameters. The specific methods are as follows: The normalized images of the leaves obtained from the four key soybean growth stages of the affected soybean plants at the flowering stage, the podding stage, the drum stage and the mature stage were further processed, and the corresponding features were extracted for each standardized image of the soybean leaves with the identified disease and insect pest types. parameters, for the leaves of soybean plants affected by the pests and diseases to be identified in a specific growth period, in order to remove the heterogeneity of the extraction results of characteristic parameters between samples, 20 plants were selected, and each plant randomly selected a standardized color image of a leaf, extracted. The image matrix R(x,y), G(x,y), B(x,y) of the red, green, and blue components of the leaf normalized color image, and then the formula F(x,y) for the three components =R(x,y)/3+G(x,y)/3+B(x,y)/3 Convert the normalized color image of the leaf into a grayscale image, perform histogram statistics on the grayscale image, and find the soybean The gray value t1 corresponding to the lowest frequency point between the two grayscale peaks of the green part of the leaf and the discolored part affected by pests and diseases The brightness value of many pixels is set to 1 to convert the grayscale image into a binary image; according to the pixel value w1 corresponding to the width of the leaf image and the measurement result of the plant corresponding to the leaf image w1' cm, calculate the value of each pixel. The actual side length z1=w1'/w1 cm, the actual area of each pixel is z1 ² cm2, according to the actual side length of the pixel z1 cm, the actual area of the pixel z1 ² cm2, and each spot is different The number of pixels corresponding to the parameter extraction result, calculate the average value of the leaf characteristic parameters of the plants affected by diseases and insect pests to be identified in this specific growth period, including the ratio a1 of the leaf spot area of the plants to be identified, and the number of leaf spots of the plants to be identified a2, to be identified The leaf spot area value of plants affected by diseases and insect pests is a3, and the perimeter value of leaf spots of plants with diseases and insect pests to be identified is a4; again, the grayscale image of plants affected by diseases and insect pests is subjected to histogram statistics to further extract the overall morphological characteristics of leaves affected by diseases and insect pests, and the soybean leaves and background colors can be found. The gray value t1' corresponding to the lowest point of the gray level frequency between the peaks of the two gray levels The value is set to 1 to realize the binarization processing of the whole soybean leaf, and extract the number of pixels corresponding to the long axis of the leaf and the number of pixels corresponding to the short axis according to the binary map of the leaf to calculate the aspect ratio of the leaf of the specific growth expectation to identify the plant disease and insect pest. a5 and the specific growth expectation to identify the leaf ellipticity of the plant with pests and diseases a6; For the roots of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters between samples, 20 standardized root images of plants were selected, and the red, green and blue components of the leaf color image were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 Convert the color leaf image to grayscale, perform histogram statistics on the root grayscale image, and find the grayscale between the two grayscale peaks of soybean root and background color The grayscale value t2 corresponding to the lowest frequency point, set the brightness value of all pixels whose grayscale is greater than the threshold t2 to 0, and set the brightness value of all pixels whose grayscale is less than t2 to 1, so as to convert the root grayscale image into two Value map, according to the pixel value w2 corresponding to the width of the root image and the measurement result of the plant corresponding to the root image w2' cm, calculate the actual side length of each pixel z2=w2'/w2 cm, the actual area of each pixel Then it is z2 ² square centimeters. According to the actual side length of the pixel z2 cm, the actual area of the pixel z2 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of the root system of plants affected by pests and diseases, the identification of the specific growth expectation is calculated. The root area value of the plant affected by diseases and insect pests is b1, and the number of root nodules of the plant to be identified is b2. Calculate the average value of the root texture characteristic parameters of the plant affected by diseases and insect pests to be identified in the specific growth period according to the grayscale map. The specific calculation process of the root texture characteristic parameters is as follows: Calculate the root contrast texture feature value b3 of the plant with plant diseases and insect pests to be identified according to formula 1, the texture feature value b4 of the root system consistency of the plants to be identified according to formula 2, the texture feature value b5 of the root system entropy value of the plants to be identified according to formula 3, and the texture feature value b5 of the plant root system to be identified according to formula 4 The eigenvalue b6 of root energy texture of plants with pests and diseases,

Among them, p(i,j) represents the value of the grayscale co-occurrence matrix calculated from the root grayscale image at the i-th row and j-th column position, and represents the pixel and gray level of the gray level i in the root grayscale image. The probability that the pixels with level j appear at the same time in a fixed direction and pixel interval, n is the difference between the two gray levels of i and j in the root gray image, and Ng represents the gray co-occurrence matrix extracted from the root gray image series; For the grains of soybean plants affected by diseases and insect pests in a specific growth period, in order to remove the heterogeneity of the extraction results of index parameters among samples, 20 plants were selected, and the red, green and blue components of the color grain image of each plant's grain were extracted. The image matrix R(x,y), G(x,y), B(x,y), for the three components according to the formula F(x,y)=R(x,y)/3+G(x, y)/3+B(x,y)/3 Convert the color grain image into a grayscale image, and perform histogram statistics on the grayscale image to find the lowest frequency between the two grayscale peaks in the soybean kernel part and the background part of the test bench The gray value t3 corresponding to the point, the brightness value of all pixels whose gray value is greater than the threshold value t3 is set to 1, and the brightness value of many pixels whose gray value is less than the threshold value t3 is set to 0, so as to convert the grayscale image of the grain into two values. Figure, calculate the actual side length of each pixel z3=w3'/w3 cm according to the pixel value w3 corresponding to the width of the grain image affected by the disease and insect pests and the measurement result of the root image corresponding to the plant w3' cm, each pixel The actual area is z3 ² square centimeters, according to the actual side length of the pixel z3 centimeters, the actual area of the pixel z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by pests and diseases, calculate the specific The average value of the grain characteristic parameters of the plants affected by diseases and insect pests to be identified during the growth period, including the number of grains of the plants to be identified c1, the value of the grain area of the plants to be identified c2, the value of the grain circumference of the plants to be identified c3, and the length of the grains of the plants to be identified. Width ratio c4, grain curvature value c5 of plants to be identified with pests and diseases, grain ellipticity value of plants to be identified with pests and diseases c6 Step 4. Establishment of database of soybean pests and diseases A database is established using the known characteristic maps of soybean plants under the influence of different types of soybean diseases and insect pests. In the database, the various characteristic parameters involved in step 3 are extracted and saved for each type of soybean plant under the influence of diseases and insect pests under different growth periods. , assign different weights to different characteristic parameters according to the difference of the types of diseases and insect pests and their effects on plants, and establish the discriminant model of the types of diseases and insect pests according to the extraction results of the characteristic parameters; Step 5. Identification and identification of soybean diseases and insect pests Using the characteristic parameters of plants affected by different types of soybean diseases and insect pests in the database under different growth periods and the discriminant model of the types of diseases and insect pests in the database to identify and identify the types of soybean diseases and insect pests that are actually affected, extract the characteristic parameters of soybean plants to be identified. The results are compared and calculated with the characteristic parameters of the plants under each disease and insect pest existing in the database. Finally, the calculation results of the discriminant model of the characteristic parameters of the soybean plants to be identified are compared with the calculation results of the discriminant model of the plants under each disease and insect pest to compare, calculate and match. So as to realize the identification and identification of soybean diseases and insect pests. 2. the effective identification method of a kind of soybean disease and insect damage according to claim 1, is characterized in that: Step 4: Use the images of soybean plants affected by all known types of diseases and insect pests in different growth stages of flowering, pod setting, drumming and maturity to establish a feature parameter database, and the impact of known diseases and insect pests in a specific growth period. The extraction and database establishment of soybean leaf characteristic parameters combined with the method of step 3 convert the color leaf image under the influence of known diseases and insect pests under the specific growth period into a grayscale image, and use the method of step 3 to perform histogram statistics on the grayscale image. , find the gray value T1 corresponding to the lowest frequency point between the two gray value peaks of the green part of the leaf and the discolored part affected by diseases and insect pests, set the brightness value of all pixels whose gray value is greater than the threshold T1 to 0, and set the gray value less than The brightness value of many pixels of the threshold T1 is set to 1, so as to convert the grayscale image into a binary image. The actual side length of the pixel Z1=W1'/W1 cm, the actual area of each pixel is Z1 ² square cm, according to the actual side length of the pixel Z1 cm, the actual area of the pixel Z1 ² square cm, and each The number of pixels corresponding to the extraction results of different parameters of each spot, and the average value of the leaf characteristic parameters of plants with known types of pests and diseases during the specific growth period is calculated, including the ratio of leaf spot area of plants with known diseases and insect pests A1, and the number of leaf spots of plants with known diseases and insect pests A2 , the leaf spot area value of known plants with diseases and insect pests is A3, and the leaf spot perimeter value of plants with known plant diseases and insect pests is A4. Perform histogram statistics to find the gray value T1' corresponding to the lowest gray level frequency between the two gray peaks of soybean leaf and background color, and set the brightness value of all pixels whose gray value is greater than the threshold T1' to 0. Set the brightness value of all pixels whose gray level is less than T1' to 1 to realize the binarization processing of the whole soybean leaf. According to the leaf binary map, the number of pixels corresponding to the long axis and the number of pixels corresponding to the short axis of the leaf are extracted and calculated. The aspect ratio value of leaf spot of plants with known plant diseases and insect pests in this specific growth period is A5 and the ellipticity value of plant spots with known plant diseases and insect pests is A6; The extraction and database establishment of soybean root characteristic parameters under the influence of known diseases and insect pests in a specific growth period are combined with the method of step 3 to convert the color root system images under the influence of known diseases and insect pests in the specific growth period into grayscale images, and the steps The method of 3 performs histogram statistics on the grayscale image, and finds the grayscale value T2 corresponding to the lowest point of the grayscale frequency between the peaks of the two grayscale parts of the soybean root system and the background color. The value is set to 0, the brightness value of all pixels whose grayscale is less than T2 is set to 1, and the root grayscale image is converted into a binary image. According to the pixel value W2 corresponding to the width of the root image and the measurement of the plant corresponding to the root image Result W2' cm, calculate the actual side length of each pixel Z2=W2'/W2 cm, the actual area of each pixel is Z2 ² square cm, according to the actual side length of the pixel Z2 cm, the actual size of the pixel The area Z2 is ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant roots affected by plant diseases and insect pests. Calculate the mean value of root characteristic parameters of all plants affected by known plant diseases and insect pests, including known plant root system area value B1, known plant diseases and insect pests The number of root nodules in the plant root system is B2. According to the formula 1 in step 3, the root contrast texture characteristic value B3 of the known plant with plant diseases and insect pests is calculated, the root consistency texture characteristic value of the plant with known plant diseases and insect pests is calculated according to the formula 2, and the texture characteristic value of the known plant root system is calculated according to the formula 3. Root entropy value texture feature value B5, calculate root energy texture feature value B6 of known plant diseases and insect pests according to formula 4; For the grains of soybean plants affected by known diseases and insect pests in a specific growth period, the color grain images under the influence of known diseases and insect pests in the specific growth period are converted into grayscale images in combination with the method in step 3, and the grayscale images are analyzed. The histogram statistics find the gray value T3 corresponding to the lowest frequency point between the two gray value peaks of the soybean grain part and the background part of the test bench, and set the brightness value of all pixels whose gray value is greater than the threshold T3 to 1, and the gray value is less than the threshold value. The brightness value of most pixels in T3 is set to 0, and the grayscale image of the grain is converted into a binary image. , calculate the actual side length of each pixel Z3=W3'/W3 cm, the actual area of each pixel is Z3 ² square cm, according to the actual side length of the pixel Z3 cm, the actual area of the pixel Z3 ² square centimeters, and the number of pixels corresponding to the extraction results of different parameters of plant grains affected by diseases and insect pests, calculate the average value of characteristic parameters of plant grains known to be affected by plant diseases and insect pests in this specific growth period, including the number of known plant grains C1, known The grain area value of plants with known diseases and insect pests is C2, the perimeter value of grains of plants with known diseases and insect pests is C3, the ratio of grain length to width of plants with known diseases and insect pests is C4, the curvature value of grains of plants with known diseases and insect pests is C5, and the ellipticity value of grains of plants with known diseases and insect pests is C6. To determine the difference in the effects _on soybean plant organs, determine the weight K ₁ of the leaf characteristic parameters of each known soybean pests and diseases, determine the weight K2 of the root characteristic parameters of each known soybean pests and diseases, and determine the weights of each known soybean pests and diseases. The weight K ₃ of the characteristic parameter of the type grain, and at the same time satisfy the normalized characteristic of the weight parameter, that is, K ₁ +K ₂ +K ₃ =1, according to the extraction results and Based on the extraction results of different organ image feature parameters of known pest species in the database, a pest and disease discrimination model is established. The discriminant model is shown in Equation 5:

3. the effective identification method of a kind of soybean disease and insect damage according to claim 2, is characterized in that: Step 5: Use the characteristic parameters of plants affected by different types of known soybean diseases and insect pests in the database in step 4 under different growth periods and the discrimination model of known types of soybean diseases and insect pests in the database to realize the identification and identification of the affected soybean diseases and insect pests. A1, a2, a3, a4, a5, a6 of leaf characteristic parameter extraction results of a plant affected by diseases and insect pests to be identified; extraction results b1, b2, b3, b4, b5, b6 of root system characteristic parameters of a plant affected by diseases and insect pests to be identified; The extraction results of grain characteristic parameters c1, c2, c3, c4, c5, and c6 of plants affected by diseases and insect pests, the extraction results of characteristic parameters of each soybean affected by diseases and insect pests to be identified and the characteristics of each known soybean affected by diseases and insect pests in this growth period. The feature parameter extraction result is brought into the discriminant model in step 4, and the P value of the model calculation result of the type of pest to be identified and all known types of pests and diseases is calculated according to the judgment model. The closer it is to the known pest and disease type corresponding to the calculation result of the P value, so as to realize the identification and identification of the soybean pest and disease type.

Images