A Korean pine growth analysis system

Abstract

The invention belongs to the field of image processing, and discloses a Korean pine growth analysis system, including a data acquisition module, a leaf length prediction module, an image segmentation module and an image recognition module; the data acquisition module is used to obtain the influencing factors in the growth process of the Korean pine; the leaf length prediction module is used to predict the leaf length of the Korean pine; the image segmentation module is used to divide the image containing the leaves of the Korean pine into multiple sub-images, and obtain the leaf area in each sub-image respectively; the image recognition module is used to identify the leaf area and judge whether the leaf area grows normally. The size of the sub-image of the present invention is not specified in advance, but is obtained based on a prediction model. Therefore, the size of the sub-image of the present invention can better adapt to Korean pines with different growth rates, which is conducive to obtaining a more reasonable sub-image size, thereby improving the accuracy of the obtained leaf area.

Description

Technical Field

[0001] The invention relates to the field of image processing, and in particular to a Korean pine growth condition analysis system. Background Art

[0002] During the growth of Korean pine, it is usually necessary to regularly analyze the growth status in order to promptly discover problems in the growth of Korean pine, maintain the healthy growth of Korean pine, and obtain better economic benefits. In the prior art, in order to analyze the growth status of Korean pine trees in a local area with similar tree ages, drones are usually used to sample and photograph some Korean pine trees in the local area, obtain images of different parts of the Korean pine, and then analyze these images to evaluate the overall growth status of the Korean pine trees in the local area.

[0003] In the process of analyzing the image of the leaf position of Korean pine, since the area of ​​a single leaf accounts for a relatively small proportion of the total leaf area, in order to improve the accuracy of the leaf area obtained, it is generally necessary to first divide the image into multiple sub-images, and then identify the leaf area in each sub-image, and then extract features from the leaf area to determine whether the leaf grows normally. However, in the prior art, for Korean pine in different areas in different local areas, they are usually directly divided into sub-images of the same size set in advance, which easily leads to the low accuracy of obtaining the leaf area. When performing image recognition on the leaves of Korean pine in an area with a faster growth rate, the complete probability of the leaf area will be reduced, because the possibility that the leaves of Korean pine are larger than the area of ​​the sub-image increases, thereby increasing the possibility that the obtained sub-image does not contain a complete leaf area. When performing image recognition on the leaves of Korean pine in an area with a slower growth rate, due to the large area of ​​the sub-image, there are more leaves in the same sub-image. When the leaves are close to each other, two leaves may be mistakenly identified as one leaf, resulting in a larger area of ​​the segmented leaf area, affecting the accuracy of the obtained leaf area.

[0004] If the accuracy of the obtained leaf region is not high enough, it will affect the accuracy of the analysis results obtained by subsequent analysis of the leaf region. Summary of the invention

[0005] The purpose of the present invention is to disclose a Korean pine growth analysis system to solve the problem of how to reasonably determine the size of the sub-image in the process of analyzing the growth of Korean pine using image recognition technology, so as to improve the accuracy of the obtained leaf area and further improve the accuracy of the analysis results obtained by analyzing the leaf area.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention provides a Korean pine growth analysis system, comprising a data acquisition module, a leaf length prediction module, an image segmentation module and an image recognition module;

[0008] The data acquisition module is used to obtain the influencing factors in the growth process of Korean pine;

[0009] The leaf length prediction module is used to obtain a set of influencing factors according to the shooting time of the image containing the leaves of the Korean pine, and predict the leaf length of the Korean pine based on the set of influencing factors;

[0010] The image segmentation module is used to divide the image containing the leaves of the red pine into a plurality of sub-images based on the length of the leaves of the red pine, and obtain the leaf area in each sub-image respectively;

[0011] The image recognition module is used to identify the leaf area and determine whether the leaf area is growing normally;

[0012] Among them, the combination of factors that affect the prediction of leaf length of Korean pine includes:

[0013] Obtain the corresponding pre-trained prediction model according to the age of Korean pine;

[0014] The set of influencing factors is input into the prediction model for calculation, and the prediction model outputs the leaf length of the Korean pine;

[0015] The image containing the leaves of the red pine is divided into a plurality of sub-images based on the length of the leaves of the red pine, including:

[0016] Calculate the actual length L1 corresponding to a pixel point in the image containing the leaves of the Korean pine;

[0017] L2 represents the leaf length of Korean pine output by the prediction model;

[0018] Then the number of rows and columns of pixels in the sub-image are ;

[0019] The image containing the leaves of Korean pine is divided into multiple pixel points with the number of rows and columns being sub-image of .

[0020] Preferably, the influencing factors include the age of the Korean pine, the temperature of the location where the Korean pine is located, the rainfall, the humidity, the light intensity, the soil moisture and the soil pH.

[0021] Preferably, the factors affecting the growth of Korean pine are obtained, including:

[0022] Starting from the germination of Korean pine, a fixed time period is used to obtain the influencing factors of the location of Korean pine.

[0023] Preferably, a set of influencing factors is obtained according to the shooting time of the image containing the leaves of Korean pine, including:

[0024] Let t1 represent the time when the red pine sprouts, and let t2 represent the time when the image containing the leaves of the red pine is taken. In the time interval [t1, t2], the influencing factors of the red pine are stored in the set of influencing factors.

[0025] Preferably, the prediction model comprises a support vector regression model.

[0026] Preferably, the training process of the support vector regression model includes:

[0027] The support vector regression models corresponding to Korean pine trees of different ages are trained separately.

[0028] Preferably, the support vector regression models corresponding to Korean pine trees of different ages are trained respectively, including:

[0029] For a Korean pine tree with an age of k, the training process of the corresponding support vector regression model is as follows:

[0030] Obtain a training data set, the training data set includes a plurality of first data, each of which includes a set of influencing factors of a red pine with an age of k and a leaf length of the red pine with an age of k; the set of influencing factors of the red pine includes influencing factors obtained during the growth process of the red pine from the budding of the red pine to the age of k;

[0031] Acquire data features of various types of influencing factors in each piece of first data, and combine the data features of various types of influencing factors in the first data and the leaf length of Korean pine to form second data;

[0032] Divide all the second data into training set and test set;

[0033] The training set is divided into multiple batches of data, and the support vector regression model is trained using the data of each batch in turn. After each batch of training is completed, the prediction performance of the support vector regression model is tested using the test set. When the prediction performance meets the set requirements, the training ends;

[0034] When the prediction performance does not meet the set requirements, the next batch of data is used to train the support vector regression model.

[0035] Preferably, the prediction performance of the support vector regression model is tested using a test set, including:

[0036] Input all the second data in the test set into the support vector regression model for prediction to obtain the predicted length;

[0037] The average absolute percentage error is calculated based on the predicted lengths corresponding to all the second data and the leaf lengths of the red pine in the second data.

[0038] Preferably, when the mean absolute percentage error is less than a set mean absolute percentage error threshold, it indicates that the prediction performance meets the set requirements.

[0039] Preferably, the data acquisition module is also used to acquire an image of a trunk region containing Korean pine;

[0040] The image segmentation module is used to obtain the trunk region from the image containing the trunk region of the red pine;

[0041] The image recognition module is used to identify the trunk area and determine whether the trunk area is growing normally.

[0042] Beneficial effects;

[0043] Compared with the prior art, in the process of acquiring the leaf area of ​​Korean pine, the present invention first acquires the leaf length through a prediction model, then determines the size of the sub-image based on the leaf length, then acquires the leaf area based on the size of the sub-image, and finally identifies the leaf area to determine whether the leaf grows normally. Since the size of the sub-image is not specified in advance but is obtained based on the prediction model, the size of the sub-image of the present invention can better adapt to Korean pine with different growth rates, which is conducive to obtaining a more reasonable sub-image size, thereby improving the accuracy of the obtained leaf area. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0045] Figure 1 The present invention is a schematic diagram of a Korean pine growth analysis system.

[0046] Figure 2 A schematic diagram of the process of acquiring a leaf region in a sub-image according to the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] like Figure 1 As shown, the present invention provides a Korean pine growth analysis system, including a data acquisition module, a leaf length prediction module, an image segmentation module and an image recognition module;

[0049] The data acquisition module is used to obtain the influencing factors in the growth process of Korean pine;

[0050] The leaf length prediction module is used to obtain a set of influencing factors according to the shooting time of the image containing the leaves of the Korean pine, and predict the leaf length of the Korean pine based on the set of influencing factors;

[0051] The image segmentation module is used to divide the image containing the leaves of the red pine into a plurality of sub-images based on the length of the leaves of the red pine, and obtain the leaf area in each sub-image respectively;

[0052] The image recognition module is used to identify the leaf area and determine whether the leaf area is growing normally;

[0053] Among them, the combination of factors that affect the prediction of leaf length of Korean pine includes:

[0054] Obtain the corresponding pre-trained prediction model according to the age of Korean pine;

[0055] The set of influencing factors is input into the prediction model for calculation, and the prediction model outputs the leaf length of the Korean pine;

[0056] The image containing the leaves of the red pine is divided into a plurality of sub-images based on the length of the leaves of the red pine, including:

[0057] Calculate the actual length L1 corresponding to a pixel point in the image containing the leaves of the Korean pine;

[0058] L2 represents the leaf length of Korean pine output by the prediction model;

[0059] Then the number of rows and columns of pixels in the sub-image are ;

[0060] The image containing the leaves of Korean pine is divided into multiple pixel points with the number of rows and columns being sub-image of .

[0061] In the process of acquiring the leaf region of Korean pine, the leaf length is first acquired through the prediction model, and then the size of the sub-image is determined based on the leaf length, and then the leaf region is acquired based on the size of the sub-image, and finally the leaf region is identified to determine whether the leaf grows normally. Since the size of the sub-image is not specified in advance, but is obtained based on the prediction model, the size of the sub-image of the present invention can better adapt to Korean pine with different growth rates, which is conducive to obtaining a more reasonable sub-image size, thereby improving the accuracy of the obtained leaf region.

[0062] The present invention can improve the accuracy of the results obtained by analyzing the leaf area of ​​the sampled Korean pine when analyzing the growth conditions of the Korean pine in the area of ​​the tree age node.

[0063] Preferably, identifying the leaf region and determining whether the leaf region is growing normally includes:

[0064] The leaf area is identified to determine whether the leaf area contains a preset type of disease. If not, it means that the leaf area is growing normally.

[0065] Preferably, the predetermined type of disease includes any one of pine needle rust, pine leaf blight, pine shrinkage and pine spot.

[0066] Preferably, if Figure 2 As shown, the leaf area is obtained in the sub-image, including:

[0067] Perform threshold segmentation on the sub-image to obtain a set UA of pixel points whose grayscale values ​​are greater than a first segmentation threshold;

[0068] Calculate the segmentation feature value of each pixel in the set UA respectively;

[0069] Obtain a set UB of pixel points whose segmentation feature values ​​in UA are greater than the set comparison value of the segmentation feature value;

[0070] Perform threshold segmentation again based on each pixel point in UB to obtain a set UC of pixel points whose grayscale values ​​are greater than the second segmentation threshold value;

[0071] Get the union UD of UA and UC;

[0072] Among all the connected domains composed of the pixels in the UD, the connected domain with the largest area is taken as the leaf region.

[0073] In the process of obtaining the leaf area, the present invention is different from the prior art method of directly taking the segmentation result as the final segmentation result. After obtaining the segmentation result, the segmentation feature value of the pixel points in UA is calculated, and threshold segmentation is performed again based on the pixel points with larger segmentation feature values. This can reduce the probability of mistakenly identifying two leaves as the same area in a single segmentation, which is conducive to obtaining a leaf area with better segmentation accuracy.

[0074] The above process is different from the method of using two different segmentation algorithms to obtain segmentation results to improve segmentation accuracy. When performing the second segmentation, the present invention only needs to calculate part of the pixel points in UA, thereby greatly reducing the number of pixel points involved in the second segmentation and effectively improving the segmentation efficiency.

[0075] By taking the largest connected domain as the leaf region, it is helpful to increase the probability of obtaining a complete leaf region.

[0076] Preferably, performing threshold segmentation on the sub-image to obtain a set UA of pixels whose grayscale values ​​are greater than the segmentation threshold includes:

[0077] Grayscale the sub-image to obtain a sub-grayscale image;

[0078] Use the Otsu algorithm to calculate the first segmentation threshold of the sub-grayscale image;

[0079] The pixel points in the sub-grayscale image whose grayscale values ​​are greater than the first segmentation threshold are stored in the set UA.

[0080] Preferably, the calculation formula for the segmentation feature value is:

[0081]

[0082] in, represents the segmentation feature value of pixel b, It means that the pixel b is the center and the side length is The number of pixels in the set dpU of pixels that do not belong to UA in the area, It means that the pixel b is the center and the side length is The total number of pixels in the area Indicates the connectivity value. If there are pixels with the same horizontal coordinate or vertical coordinate in dpU, then The value of is 1, otherwise, The value is 0; neipd means that the pixel b is the center and the side length is The set of pixels in the area Represents the image gradient of pixel i; Represents the average value of image gradient in neipd, , and Represent the first reference weight, the second reference weight and the third reference weight respectively.

[0083] The segmentation feature value of the present invention is comprehensively calculated from multiple directions, which is conducive to making the segmentation feature value better reflect the possibility of local image segmentation around the pixel point b. Specifically, when The larger the value, the more drastic the change in the image gradient of the pixel in neipd, the larger the segmentation feature value of the present invention, indicating that the possibility of local segmentation around the pixel b is greater, so that the edge of the segmentation result obtained by using the first segmentation threshold can be reconfirmed to obtain a more accurate edge of the leaf area. This is conducive to more accurately judging whether the leaves are growing normally during subsequent image recognition.

[0084] Preferably, the first reference weight, the second reference weight and the third reference weight are , and .

[0085] Preferably, the comparison value of the segmentation feature value is set to .

[0086] Preferably, threshold segmentation is performed again based on each pixel point in UB to obtain a set UC of pixels whose grayscale values ​​are greater than the segmentation threshold, including:

[0087] Each pixel in UB is processed as follows:

[0088] Let ub represent the pixel point in UB. In the sub-grayscale image, get the pixel point with radius centered on ub. The circular area of

[0089] Use the Otsu algorithm to calculate the second segmentation threshold of the circular area;

[0090] In the second segmentation threshold, the pixel points whose grayscale values ​​are greater than the segmentation threshold are stored in the set UC.

[0091] Preferably, The calculation formula is:

[0092]

[0093] It means that the size is centered on ub. In the area, the number of pixels whose segmentation feature value is greater than the control value of the set segmentation feature value, subs represents the pixel with ub as the center and the size is The set of pixel points in the region where the segmentation feature value is greater than the control value of the set segmentation feature value, Represents the sum of the distances between j and each pixel in subs except j. , max means getting the maximum value in the set, bcR means a preset integer, Represents a weighted ratio value.

[0094] The radius of the circular area of ​​the present invention is not fixed, but is determined by The value of and the average of the distances between the pixels in subs are weighted. The larger the value, the smaller the sum of the distances between the pixels in subs, and the smaller the radius. In this way, there can be more pixels in UB around ub, and the denser the distribution of these pixels, the smaller the radius, which is conducive to reducing the radius to reduce interference and reduce the degree of impact on the accuracy of segmentation. When there are smaller pixels in UB around ub and the distribution of these pixels is sparse, the larger the radius value is, which is conducive to referring to more pixels and improving the accuracy of segmentation.

[0095] Preferably, the weighted ratio value is 0.5.

[0096] Preferably, the preset integer is 20.

[0097] Preferably, the value of H is 11.

[0098] Preferably, the influencing factors include the age of the Korean pine, the temperature of the location where the Korean pine is located, the rainfall, the humidity, the light intensity, the soil moisture and the soil pH.

[0099] The temperature range for the growth of Korean pine is usually 10°C to 20°C. Extreme high or low temperatures can have a negative impact on its growth. Low temperatures may cause frost damage, affecting the growth and survival rate of Korean pine. Korean pine has a high demand for water, but it also requires well-drained soil. Too much precipitation can cause root rot, while too little will affect growth. The seasonal distribution of precipitation can also affect the growth of Korean pine, especially the water supply during the dry season. Appropriate air humidity can help reduce the occurrence of diseases, but too high humidity may lead to an increase in fungal diseases. Light intensity: Korean pine usually requires sufficient sunlight to carry out photosynthesis and promote growth. Shaded environment will inhibit its growth. Daylight duration: Sufficient sunshine time helps to improve the growth rate and health level of Korean pine.

[0100] Preferably, the factors affecting the growth of Korean pine are obtained, including:

[0101] Starting from the germination of Korean pine, a fixed time period is used to obtain the influencing factors of the location of Korean pine.

[0102] For example, the influencing factors may be obtained with a time period of one month.

[0103] Preferably, a set of influencing factors is obtained according to the shooting time of the image containing the leaves of Korean pine, including:

[0104] Let t1 represent the time when the red pine sprouts, and let t2 represent the time when the image containing the leaves of the red pine is taken. In the time interval [t1, t2], the influencing factors of the red pine are stored in the set of influencing factors.

[0105] Preferably, the prediction model comprises a support vector regression model.

[0106] Specifically, the prediction model may also be a decision tree regression model, a random forest regression model, or the like.

[0107] Preferably, the training process of the support vector regression model includes:

[0108] The support vector regression models corresponding to Korean pine trees of different ages are trained separately.

[0109] Since there are differences in the length of leaves of Korean pines of different ages, the present invention can better obtain data features corresponding to the influencing factors of Korean pines of different ages by training different prediction models of different ages, thereby improving the accuracy of prediction.

[0110] Preferably, the support vector regression models corresponding to Korean pine trees of different ages are trained respectively, including:

[0111] For a Korean pine tree with an age of k, the training process of the corresponding support vector regression model is as follows:

[0112] Obtain a training data set, the training data set includes a plurality of first data, each of which includes a set of influencing factors of a red pine with an age of k and a leaf length of the red pine with an age of k; the set of influencing factors of the red pine includes influencing factors obtained during the growth process of the red pine from the budding of the red pine to the age of k;

[0113] Acquire data features of various types of influencing factors in each piece of first data, and combine the data features of various types of influencing factors in the first data and the leaf length of Korean pine to form second data;

[0114] Divide all the second data into training set and test set;

[0115] The training set is divided into multiple batches of data, and the support vector regression model is trained using the data of each batch in turn. After each batch of training is completed, the prediction performance of the support vector regression model is tested using the test set. When the prediction performance meets the set requirements, the training ends;

[0116] When the prediction performance does not meet the set requirements, the next batch of data is used to train the support vector regression model.

[0117] Preferably, each batch of data may include 32 pieces of second data.

[0118] Specifically, data features include mean, variance, maximum, minimum and sum.

[0119] During each training, the gradient of the loss function with respect to the model parameters is calculated by backpropagation, and the model parameters are updated according to the gradient using an optimization algorithm (such as gradient descent).

[0120] Preferably, the prediction performance of the support vector regression model is tested using a test set, including:

[0121] Input all the second data in the test set into the support vector regression model for prediction to obtain the predicted length;

[0122] The average absolute percentage error is calculated based on the predicted lengths corresponding to all the second data and the leaf lengths of the red pine in the second data.

[0123] Preferably, when the mean absolute percentage error is less than a set mean absolute percentage error threshold, it indicates that the prediction performance meets the set requirements.

[0124] Preferably, the mean absolute percentage error threshold is 5%.

[0125] Preferably, the data acquisition module is also used to acquire an image of a trunk region containing Korean pine;

[0126] The image segmentation module is used to obtain the trunk region from the image containing the trunk region of the red pine;

[0127] The image recognition module is used to identify the trunk area and determine whether the trunk area is growing normally.

[0128] Specifically, the trunk area is identified to determine whether the trunk area is growing normally, including:

[0129] Determine whether the trunk area has any of the diseases including pine trunk rot, pine knot disease, and bark cancer. If not, it means that the trunk area is growing normally.

[0130] It is obvious to those skilled in the art that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention; therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, and therefore it is intended to include all changes within the meaning and scope of the equivalent elements of the claims within the present invention, and any figure mark in the claims should not be regarded as limiting the claims involved.

[0131] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A Korean pine growth analysis system, characterized in that: It includes a data acquisition module, a leaf length prediction module, an image segmentation module and an image recognition module; The data acquisition module is used to obtain the influencing factors in the growth process of Korean pine; The leaf length prediction module is used to obtain a set of influencing factors according to the shooting time of the image containing the leaves of the Korean pine, and predict the leaf length of the Korean pine based on the set of influencing factors; The image segmentation module is used to divide the image containing the leaves of the red pine into a plurality of sub-images based on the length of the leaves of the red pine, and obtain the leaf area in each sub-image respectively; The image recognition module is used to identify the leaf area and determine whether the leaf area is growing normally; Among them, the combination of factors that affect the prediction of leaf length of Korean pine includes: Obtain the corresponding pre-trained prediction model according to the age of Korean pine; The set of influencing factors is input into the prediction model for calculation, and the prediction model outputs the leaf length of the Korean pine; The image containing the leaves of the red pine is divided into a plurality of sub-images based on the length of the leaves of the red pine, including: Calculate the actual length L1 corresponding to a pixel point in the image containing the leaves of the Korean pine; L2 represents the leaf length of Korean pine output by the prediction model; Then the number of rows and columns of pixels in the sub-image are The image containing the leaves of Korean pine is divided into multiple pixel points with the number of rows and columns being sub-image of Get the leaf area in the sub-image, including: Perform threshold segmentation on the sub-image to obtain a set UA of pixel points whose grayscale values ​​are greater than a first segmentation threshold; Calculate the segmentation feature value of each pixel in the set UA respectively; Obtain a set UB of pixel points whose segmentation feature values ​​in UA are greater than the set comparison value of the segmentation feature value; Perform threshold segmentation again based on each pixel point in UB to obtain a set UC of pixel points whose grayscale values ​​are greater than the second segmentation threshold value; Get the union UD of UA and UC; Among all the connected domains composed of the pixels in the UD, the connected domain with the largest area is regarded as the leaf area; Perform threshold segmentation on the sub-image to obtain the set UA of pixels whose grayscale values ​​are greater than the segmentation threshold, including: Grayscale the sub-image to obtain a sub-grayscale image; Use the Otsu algorithm to calculate the first segmentation threshold of the sub-grayscale image; In the sub-grayscale image, the pixels whose grayscale values ​​are greater than the first segmentation threshold are stored in the set UA; The calculation formula of the segmentation feature value is: Among them, sfeavl b Represents the segmentation feature value of pixel b, ndbt b Indicates the number of pixels in the set dpU of pixels that do not belong to UA in the area with a side length of 3×3 and centered at pixel b, nall b Represents the total number of pixels in an area with a side length of 3×3 centered on pixel b, psc b Indicates the connectivity value. If there are pixels with the same horizontal coordinate or vertical coordinate in dpU, psc b The value of is 1, otherwise, psc b The value is 0; neipd represents the set of pixels in the area with a side length of 3×3 and centered at pixel b, grys i represents the image gradient of pixel i; grysave represents the average value of the image gradient in neipd, λ1, λ2 and λ3 represent the first reference weight, the second reference weight and the third reference weight respectively; Perform threshold segmentation again based on each pixel in UB to obtain a set UC of pixels whose grayscale values ​​are greater than the segmentation threshold, including: Each pixel in UB is processed as follows: Let ub represent the pixel in UB. In the sub-grayscale image, get the pixel with radius R centered on ub. ub Circular area of Use the Otsu algorithm to calculate the second segmentation threshold of the circular area; In the second segmentation threshold, the pixel points whose grayscale value is greater than the segmentation threshold are stored in the set UC; R ub The calculation formula is: nbg represents the number of pixels whose segmentation feature values ​​are greater than the control value of the set segmentation feature value in the H×H area centered on ub, subs represents the set of pixels whose segmentation feature values ​​are greater than the control value of the set segmentation feature value in the H×H area centered on ub, alldist j Represents the sum of the distances between j and each pixel in subs except j. max means obtaining the maximum value in the set, bcR means a preset integer, and γ means a weighted ratio value.

2. A Korean pine growth analysis system according to claim 1, characterized in that: Influencing factors include the age of the red pine, the temperature of the location where the red pine is located, rainfall, humidity, light intensity, soil moisture and soil pH.

3. A Korean pine growth analysis system according to claim 1, characterized in that: Obtain the factors affecting the growth of red pine, including: Starting from the germination of Korean pine, a fixed time period is used to obtain the influencing factors of the location of Korean pine.

4. A Korean pine growth analysis system according to claim 1, characterized in that: A set of influencing factors is obtained according to the shooting time of the image containing the leaves of the red pine, including: Let t1 represent the time when the red pine sprouts, and let t2 represent the time when the image containing the leaves of the red pine is taken. In the time interval [t1, t2], the influencing factors of the red pine are stored in the set of influencing factors.

5. The Korean pine growth analysis system according to claim 1, characterized in that: The prediction models include support vector regression models.

6. A Korean pine growth analysis system according to claim 5, characterized in that: The training process of the support vector regression model includes: The support vector regression models corresponding to Korean pine trees of different ages are trained separately.

7. A Korean pine growth analysis system according to claim 6, characterized in that: The support vector regression models corresponding to Korean pine trees of different ages are trained separately, including: For a Korean pine tree with an age of k, the training process of the corresponding support vector regression model is as follows: Obtain a training data set, the training data set includes a plurality of first data, each of which includes a set of influencing factors of a red pine with an age of k and a leaf length of the red pine with an age of k; the set of influencing factors of the red pine includes influencing factors obtained during the growth process of the red pine from the budding of the red pine to the age of k; Acquire data features of various types of influencing factors in each piece of first data, and combine the data features of various types of influencing factors in the first data and the leaf length of Korean pine to form second data; Divide all the second data into training set and test set; The training set is divided into multiple batches of data, and the support vector regression model is trained using the data of each batch in turn. After each batch of training is completed, the prediction performance of the support vector regression model is tested using the test set. When the prediction performance meets the set requirements, the training ends; When the prediction performance does not meet the set requirements, the next batch of data is used to train the support vector regression model.

8. A Korean pine growth analysis system according to claim 7, characterized in that: Use the test set to test the predictive performance of the support vector regression model, including: Input all the second data in the test set into the support vector regression model for prediction to obtain the predicted length; The average absolute percentage error is calculated based on the predicted lengths corresponding to all the second data and the leaf lengths of the red pine in the second data.

9. A Korean pine growth analysis system according to claim 8, characterized in that: When the mean absolute percentage error is less than the set mean absolute percentage error threshold, it means that the prediction performance meets the set requirements.

10. The Korean pine growth analysis system according to claim 8, characterized in that: The data acquisition module is also used to acquire an image of a trunk region containing red pine; The image segmentation module is used to obtain the trunk region from the image containing the trunk region of the red pine; The image recognition module is used to identify the trunk area and determine whether the trunk area is growing normally.

Images