A leaf image pattern classification method and system for panicle fertilization of double-cropping rice

Abstract

The application provides a leaf image pattern classification method and system for double-cropping rice ear fertilization, and the method comprises the following steps: in a preset time window of target double-cropping rice, first leaf image parameter arrays and second leaf image parameter arrays are obtained by collecting a plurality of positions of the first leaf and the second leaf; image parameter discrete analysis is performed on the parameter arrays to obtain a first discrete degree and a second discrete degree, and a leaf contamination degree is calculated by combining a preset discrete degree; average processing and leaf pattern classification are performed on the parameter arrays, two leaf pattern arrays are obtained based on the image parameters and the contamination degree respectively, and suitable leaf patterns are determined to perform double-cropping rice ear fertilization. The application solves the technical problem that the accuracy of fertilization recommendation is not high in different fields due to insufficient test data in the prior art, and achieves the technical effects of realizing precise regulation of ear fertilization and improving the accuracy of fertilization recommendation by analyzing the pattern classification of double-cropping rice leaf image parameters and contamination degrees.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image processing, and in particular to a leaf image pattern classification method and system for double-cropping rice panicle fertilization. BACKGROUND

[0002] At present, the commonly used fertilization recommendation method is mainly based on multi-year field test data, and the fertilization amount of crops is calculated through yield response and agronomic efficiency model. However, due to the problem of insufficient test data, the application effect of these methods on different plots is obviously different, which affects the accuracy of fertilization recommendation, and it is difficult to meet the requirements of precision agriculture of farmers on fertilization management. Especially for double-cropping rice, due to the difference of early and late rice growth environment and nutrient demand, the fertilization amount calculated by the recommended fertilization method is often difficult to adapt to the actual nutrient status of the plot, resulting in improper fertilization and affecting yield and quality. SUMMARY

[0003] The present application aims at the technical problem of poor accuracy of fertilization recommendation in different plots due to insufficient test data in the prior art, and provides a leaf image pattern classification method and system for double-cropping rice panicle fertilization to solve the problem on the basis of the recommended fertilization method, so as to improve the precision of the recommended fertilization method.

[0004] The technical solution of the present application to solve the above technical problems is as follows:

[0005] In a first aspect, the present application provides a leaf image pattern classification method for double-cropping rice panicle fertilization, comprising: in a preset time window of a target double-cropping rice, acquiring a first leaf image parameter array and a second leaf image parameter array at multiple positions of a first leaf and a second leaf; performing image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discrete degree and a second discrete degree, and combining a preset first discrete degree and a preset second discrete degree to calculate a first leaf contamination degree and a second leaf contamination degree; performing average processing and leaf pattern classification on the first leaf image parameter array and the second leaf image parameter array to obtain a first leaf pattern array, and performing leaf pattern classification according to the first leaf contamination degree and the second leaf contamination degree to obtain a second leaf pattern array; and processing a leaf pattern according to the first leaf pattern array and the second leaf pattern array to perform double-cropping rice panicle fertilization.

[0006] In a second aspect, the present application provides a leaf image pattern classification system for double-cropping rice panicle fertilization, comprising: a leaf collection module for collecting a first leaf image parameter array and a second leaf image parameter array at a plurality of positions of a first leaf and a second leaf of a target double-cropping rice in a preset time window; a pollution degree analysis module for performing image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discrete degree and a second discrete degree, combining a preset first discrete degree and a preset second discrete degree, and calculating a first leaf pollution degree and a second leaf pollution degree; a leaf pattern classification module for performing averaging processing and leaf pattern classification on the first leaf image parameter array and the second leaf image parameter array to obtain a first leaf pattern array, performing leaf pattern classification according to the first leaf pollution degree and the second leaf pollution degree, and obtaining a second leaf pattern array; and a fertilization decision module for processing a leaf pattern according to the first leaf pattern array and the second leaf pattern array to perform double-cropping rice panicle fertilization.

[0007] The present application has the following advantages:

[0008] In a preset time window of a target double-cropping rice, a first leaf image parameter array and a second leaf image parameter array are collected at a plurality of positions of a first leaf and a second leaf, so that SPAD values (chlorophyll content parameters) of a plurality of positions of the third leaf (the first leaf) and the fourth leaf (the second leaf) are collected at the key time point of the second leaf tip of the double-cropping rice, and a parameter array is formed to provide basic data for subsequent analysis. Since the SPAD values at different positions of the leaf (such as the leaf tip and the leaf base) naturally differ, a plurality of positions need to be collected to form an array. The first leaf image parameter array and the second leaf image parameter array are subjected to image parameter discrete analysis to obtain a first discrete degree and a second discrete degree, and a first leaf pollution degree and a second leaf pollution degree are calculated by combining a preset first discrete degree and a preset second discrete degree. By analyzing the discrete degree of the SPAD values of the two leaves and comparing it with the preset normal leaf discrete degree (such as the difference between the SPAD values of the middle part and the edge of the leaf under normal circumstances is between 10% and 15%), the pollution degree of the leaf is calculated. When the leaf surface of the double-cropping rice is polluted by dust and other external substances, it will interfere with the accurate measurement of the SPAD instrument on the chlorophyll content, resulting in deviation of the measured value. Dust and other pollutants covering the surface of the leaf will cause the SPAD value measurement result to be low, and the pollution degree at different positions of the leaf is different, resulting in an increase in the discrete degree of the SPAD value.

[0009] The first leaf image parameter array and the second leaf image parameter array are subjected to averaging processing and leaf pattern classification to obtain a first leaf pattern array, and according to the first leaf contamination degree and the second leaf contamination degree, the leaf pattern classification is performed to obtain a second leaf pattern array. When the ratio of the first leaf image parameter to the second leaf image parameter is greater than 1, it indicates that nitrogen is being transferred to the third leaf (new leaf), and it is necessary to apply or increase the amount of panicle fertilizer; when the ratio is less than 1, it indicates that the SPAD value of the fourth leaf is higher than that of the third leaf, indicating that the nitrogen nutrition is sufficient or excessive, and it is necessary to reduce or not apply panicle fertilizer. At the same time, by analyzing the ratio of the first leaf contamination degree to the second leaf contamination degree, when the ratio is greater than 1, it indicates that the contamination degree of the third leaf is higher than that of the fourth leaf, indicating that the third leaf is more prone to adsorb dust, the surface is rougher, and the nitrogen content is lower than that of the fourth leaf, while the nitrogen of the fourth leaf is excessive, the transfer ratio to the third leaf is low, and the leaf appears to be greedy green, and it is necessary to reduce or not apply panicle fertilizer; when the ratio is less than 1, it indicates that the contamination degree of the third leaf is lower than that of the fourth leaf, indicating that the third leaf is not prone to adsorb dust, the nitrogen nutrition is in good condition, the nitrogen of the fourth leaf is normally transferred to the third leaf, and the yellowing phenomenon appears. It is necessary to apply panicle fertilizer to meet the growth needs. According to the first leaf pattern array and the second leaf pattern array, the leaf pattern is obtained by processing, the panicle fertilizer of the double-cropping rice is applied, the mode classification results are comprehensively taken as the final leaf pattern according to the mode classification results, and the amount of panicle fertilizer of the double-cropping rice is determined. According to different leaf patterns, it can be determined to normally apply fertilizer, or to increase the amount of fertilizer, or to reduce the amount of fertilizer, or even not to apply fertilizer, so as to realize precise regulation and control of fertilization.

[0010] Through the above technical solution, the present application overcomes the technical problem in the prior art that the accuracy of fertilization recommendation is insufficient on different plots due to insufficient test data, and realizes the technical effect of precise regulation and control of panicle fertilizer and improvement of the accuracy of fertilization recommendation by analyzing the mode classification of leaf image parameters and contamination degrees of double-cropping rice. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 A flowchart of a leaf image mode classification method for panicle fertilization of double-cropping rice is provided.

[0012] Figure 2 A structure diagram of a leaf image mode classification system for panicle fertilization of double-cropping rice is provided.

[0013] In the drawings, the components represented by the numbers are as follows:

[0014] The leaf collecting module 11, the contamination degree analysis module 12, the mode classification module 13, and the fertilization decision module 14. DETAILED DESCRIPTION

[0015] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0016] In the description of the present application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise explicitly specified.

[0017] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can realize the present application without using these specific details. In other examples, well-known structures and processes will not be described in detail in order to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope of principles and features disclosed.

[0018] Embodiment one, as shown in the present application, provides a leaf image pattern classification method for double-cropping rice panicle fertilization, which comprises: Figure 1

[0019] S100: In a preset time window of the target double-cropping rice, a first leaf image parameter array and a second leaf image parameter array are obtained at multiple positions of the first leaf and the second leaf.

[0020] ​Specifically, when the target double-cropping rice grows to a certain stage (i.e., a preset time window), the first leaf and the second leaf of the double-cropping rice plant are selected, and the SPAD values are measured at multiple different positions of the first leaf and the second leaf respectively by using a SPAD instrument to obtain a set of leaf image parameters (SPAD values) respectively, thereby forming a first leaf image parameter array and a second leaf image parameter array. The SPAD value represents a non-destructive measurement index of chlorophyll content and can reflect the nitrogen nutrition status of the plant. The higher the leaf SPAD value, the higher the chlorophyll content, which means the better the nitrogen nutrition status of the plant. The first leaf image parameter array refers to an ordered data set composed of SPAD values measured at multiple positions of the first leaf (the third leaf from the top), which can be represented as {SPAD 11 , SPAD 12 ,..., SPAD 1n}, wherein each SPAD value corresponds to a specific position on the leaf. Similarly, the second leaf image parameter array refers to an ordered data set composed of SPAD values measured at multiple positions of the second leaf (the fourth leaf from the top), which can be represented as {SPAD 21 , SPAD 22 ,..., SPAD 2m}.

[0021] The first leaf refers to the third leaf from the top of the double-cropping rice, and the second leaf refers to the fourth leaf from the top of the double-cropping rice. The physiological state of these two leaves can reflect the nitrogen nutrition status of the rice. The multiple positions refer to selecting different sampling points on the same leaf, such as the base, middle and top of the leaf, or different regions such as the vicinity of the main vein of the leaf and the edge of the leaf, thereby obtaining comprehensive information on the nutritional status of the leaf and avoiding errors that may be caused by single-point measurement, providing more reliable basic data for subsequent leaf pattern classification.

[0022] By collecting the image parameter arrays of the first leaf and the second leaf in the preset time window, basic data is provided for subsequent analysis of the nitrogen nutrition status of the double-cropping rice and determination of the application strategy of the panicle fertilizer.

[0023] S200: Perform image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discrete degree and a second discrete degree, and combine a preset first discrete degree and a preset second discrete degree to calculate and obtain a first leaf contamination degree and a second leaf contamination degree.

[0024] Specifically, first, the first leaf image parameter array and the second leaf image parameter array obtained by collection are respectively subjected to discrete analysis calculation, to obtain a first dispersion and a second dispersion reflecting the data dispersion degree. The discrete analysis is realized by calculating the standard deviation of the plurality of leaf image parameters in the first leaf image parameter array and the second leaf image parameter array, specifically by calculating the standard deviation of the plurality of SPAD values in the first leaf image parameter array and the second leaf image parameter array respectively, to obtain the first dispersion and the second dispersion. The greater the standard deviation value, the greater the difference between the SPAD values at different positions on the leaf, and the higher the dispersion degree.

[0025] Then, the calculated first dispersion is compared with a preset first dispersion, the amplitude of the first dispersion deviating from the preset first dispersion is calculated, and a first leaf contamination degree is obtained. At the same time, the second dispersion is compared with a preset second dispersion, the amplitude of the second dispersion deviating from the preset second dispersion is calculated, and a second leaf contamination degree is obtained. The preset first dispersion refers to the standard dispersion value that the first leaf (the third leaf from the bottom) should have under the normal growth state without being polluted by dust and other external factors; the preset second dispersion refers to the standard dispersion value that the second leaf (the fourth leaf from the bottom) should have under the normal growth state without being polluted by dust and other external factors. The first leaf contamination degree reflects the roughness degree and dust adsorption of the first leaf surface caused by nitrogen deficiency. Under normal circumstances, the difference between the SPAD values of the middle part (near the main vein) and the edge of the leaf can reach 10%-15%, which is caused by the difference in structure and physiological state of different parts of the leaf. When the leaf surface becomes rough due to nitrogen deficiency, it is easy to adsorb dust, which will change this natural difference and affect the measurement result of the dispersion. The second leaf contamination degree reflects the roughness degree and dust adsorption of the second leaf surface caused by nitrogen deficiency, and its principle is the same as that of the first leaf contamination degree.

[0026] By obtaining the first leaf contamination degree and the second leaf contamination degree, the nitrogen nutrition status of the leaf can be evaluated, and a reference basis is provided for subsequent fertilization decision.

[0027] S300: The first leaf image parameter array and the second leaf image parameter array are subjected to averaging processing and leaf pattern classification, to obtain a first leaf pattern array. According to the first leaf contamination degree and the second leaf contamination degree, leaf pattern classification is performed to obtain a second leaf pattern array.

[0028] Specifically, first, the first leaf image parameter array and the second leaf image parameter array are respectively subjected to average calculation processing to obtain the first leaf image parameter and the second leaf image parameter. The average processing refers to calculating the average value of the SPAD values of multiple positions collected to obtain a single parameter that can represent the overall nitrogen status of the whole leaf. For example, the first leaf image parameter refers to the average value of all SPAD values in the first leaf image parameter array; the second leaf image parameter refers to the average value of all SPAD values in the second leaf image parameter array.

[0029] Then, the ratio of the first leaf image parameter and the second leaf image parameter is calculated, the leaf pattern classification is performed, and the first leaf pattern array is obtained. The ratio of the first leaf image parameter and the second leaf image parameter reflects the nitrogen distribution relationship between the third leaf and the fourth leaf. When the ratio is greater than 1, it indicates that nitrogen is being transferred to the new leaf (the third leaf), the plant is normal or lacks fertilizer, and normal ear fertilizer or increased ear fertilizer needs to be applied; when the ratio is less than 1, it may indicate that nitrogen is sufficient or excessive, and the amount of ear fertilizer needs to be appropriately reduced or not applied. At the same time, the ratio of the first leaf contamination degree and the second leaf contamination degree is calculated, the leaf pattern classification is performed, and the second leaf pattern array is obtained. The ratio of the first leaf contamination degree and the second leaf contamination degree reflects the difference in surface state of the third leaf and the fourth leaf caused by the difference in nitrogen nutrition. The greater the ratio, the more likely the third leaf is to adsorb dust relative to the fourth leaf, indicating that the surface of the third leaf is rough and nitrogen deficient, while the fourth leaf is nitrogen sufficient, and nitrogen is mainly concentrated in the old leaf and less transferred to the new leaf, requiring reduced or no ear fertilizer application. The smaller the ratio, the less likely the third leaf is to adsorb dust relative to the fourth leaf, indicating that the surface of the third leaf is relatively smooth and the nitrogen nutrition status is good, and nitrogen is being transferred from the old leaf to the new leaf, requiring additional fertilizer application to meet the growth needs.

[0030] By obtaining two different dimensions of leaf pattern classification based on the ratio of SPAD values and the ratio of contamination degrees, the nitrogen nutrition status of double-cropping rice can be comprehensively evaluated from multiple angles, providing a comprehensive and accurate reference basis for subsequent ear fertilizer application decision-making.

[0031] S400: According to the first leaf pattern array and the second leaf pattern array, a leaf pattern is obtained by processing, and ear fertilizer application of double-cropping rice is performed.

[0032] Specifically, first, the leaf pattern with the largest occurrence ratio is selected from the first leaf pattern array and the second leaf pattern array to obtain a fused leaf pattern. This selection method is similar to a voting mechanism, that is, the leaf pattern with the highest frequency of occurrence is found from the two pattern arrays as the final leaf pattern decision basis. Selecting the leaf pattern with the most occurrences can reduce the error caused by a single judgment and improve the reliability of the decision.

[0033] Then, according to the obtained fusion leaf pattern, the application strategy of the panicle fertilizer of the double-cropping rice is determined. For example, for the double-cropping early rice, when the fusion leaf pattern shows that the leaf color of the third leaf from the bottom is similar to that of the fourth leaf from the bottom (within ± 5% of the SPAD value), the panicle fertilizer is normally applied according to the original calculated value; when the fusion leaf pattern shows that the leaf color of the third leaf from the bottom is lighter than that of the fourth leaf from the bottom (5%-15% lower in the SPAD value), 50% of the original calculated value is applied as the panicle fertilizer; when the fusion leaf pattern shows that the leaf color of the third leaf from the bottom is light (more than 15% lower in the SPAD value), no panicle fertilizer is applied; when the fusion leaf pattern shows that the leaf color of the third leaf from the bottom is darker than that of the fourth leaf from the bottom (5%-15% higher in the SPAD value), 120% of the original calculated value is applied as the panicle fertilizer; and when the fusion leaf pattern shows that the leaf color of the third leaf from the bottom is dark (more than 15% higher in the SPAD value), 140% of the original calculated value is applied as the panicle fertilizer. The fertilization strategy of the double-cropping late rice is similar, but the specific threshold and fertilization ratio are different. The original calculated value refers to the recommended application amount of the panicle fertilizer calculated based on the soil nutrient content, target yield, planting density and other parameters.

[0034] Through the leaf pattern-based fertilization decision method, precise fertilization can be performed according to the actual nitrogen nutritional status of the double-cropping rice, the blindness that may be caused by traditional experience fertilization is avoided, the limitations that may exist in the application of the existing fertilization system in different fields are made up, the accuracy and pertinence of the application of the panicle fertilizer of the double-cropping rice are improved, and the fertilizer utilization rate and the yield of the double-cropping rice are improved.

[0035] Further, at a preset time window of the target double-cropping rice, a first leaf image parameter array and a second leaf image parameter array are obtained at multiple positions of the first leaf and the second leaf, including:

[0036] S110: at the tip-appearing stage of the second leaf from the bottom of the target double-cropping rice, multiple positions of the first leaf and the second leaf on the target double-cropping rice are selected, wherein the first leaf is the third leaf from the bottom and the second leaf is the fourth leaf from the bottom;

[0037] S120: at the multiple positions of the first leaf and the second leaf, SPAD parameters are respectively tested and collected to obtain a first leaf image parameter array and a second leaf image parameter array.

[0038] In a feasible implementation, at the tip-appearing stage of the second leaf from the bottom of the target double-cropping rice, multiple positions of the first leaf and the second leaf on the target double-cropping rice are selected, wherein the first leaf is the third leaf from the bottom and the second leaf is the fourth leaf from the bottom. The tip-appearing stage of the second leaf from the bottom refers to a specific period when the second leaf from the bottom just exposes the leaf tip from the leaf sheath in the growth process of the rice. This period is a key stage for the transition from vegetative growth to reproductive growth of the rice, and the leaf parameter collection at this stage can more accurately reflect the nitrogen nutritional status of the rice. At this stage, the third leaf from the bottom and the fourth leaf from the bottom have been fully unfolded, and their physiological characteristics are stable, which can be ideal sampling objects for evaluating the nitrogen status of the rice.

[0039] Subsequently, SPAD parameters are tested at multiple positions of the first leaf and the second leaf respectively to obtain a first leaf image parameter array and a second leaf image parameter array. Specifically, the SPAD instrument is used to measure at selected multiple positions, such as different regions of the leaf blade, such as the base, middle, top, or near the main vein and the leaf edge, and a SPAD value is obtained at each position, which collectively constitutes the first leaf image parameter array and the second leaf image parameter array. By collecting data at multiple positions, the spatial differences in nitrogen distribution of the leaf blade can be more comprehensively understood, providing more reliable basic data for subsequent dispersion analysis and pollution degree calculation.

[0040] Through the above steps, the SPAD value distribution of the third leaf and the fourth leaf of the double-cropping rice can be obtained within a suitable time window, providing basic data for subsequent dispersion analysis and leaf pattern classification.

[0041] Further, the first leaf image parameter array and the second leaf image parameter array are subjected to image parameter dispersion analysis to obtain a first dispersion and a second dispersion, including:

[0042] S210: performing image parameter dispersion analysis on the first leaf image parameter array to obtain a first dispersion, wherein the image parameter dispersion analysis is performed by calculating the standard deviation of the multiple leaf image parameters in the first leaf image parameter array;

[0043] S220: performing image parameter dispersion analysis on the second leaf image parameter array to obtain a second dispersion.

[0044] In a preferred embodiment, the image parameter dispersion analysis is performed by calculating the standard deviation of the multiple leaf image parameters in the first leaf image parameter array to obtain the first dispersion. Specifically, the SPAD values collected at multiple positions on the first leaf (the third leaf) are taken as a group of data, and the standard deviation of the group of data is calculated to obtain the first dispersion. The standard deviation can effectively reflect the dispersion degree of the data, and the larger the standard deviation, the greater the difference in SPAD values at different positions of the leaf blade, and the higher the dispersion degree. Under normal circumstances, there will be natural differences in SPAD values between the middle and the edge of the leaf blade, and this dispersion analysis can capture the degree of such differences. At the same time, the second leaf image parameter array is subjected to image parameter dispersion analysis to obtain a second dispersion. Specifically, the SPAD values collected at multiple positions on the second leaf (the fourth leaf) are calculated to obtain the standard deviation, quantifying the variation degree of the SPAD values at different positions of the second leaf, providing data basis for subsequent analysis of the surface condition of the leaf blade.

[0045] Through the above steps, the first dispersion and the second dispersion reflecting the dispersion degree of the SPAD value distribution of the first leaf and the second leaf are obtained respectively, which will be used for subsequent calculation of the pollution degree of the leaf blade, and further evaluation of the nitrogen nutrition status of the double-cropping rice.

[0046] Further, in combination with the preset first dispersion and the preset second dispersion, the first leaf pollution degree and the second leaf pollution degree are calculated and obtained, including:

[0047] S230: Obtain the first dispersion and the second dispersion of the same variety of double-cropping rice in the preset time window in the historical time, to obtain a sample first dispersion set and a sample second dispersion set;

[0048] S240: Calculate and configure the preset first dispersion and the preset second dispersion according to the sample first dispersion set and the sample second dispersion set;

[0049] S250: Calculate the amplitudes of the first dispersion and the second dispersion deviating from the preset first dispersion and the preset second dispersion respectively, to obtain the first leaf pollution degree and the second leaf pollution degree.

[0050] Specifically, first, the SPAD value dispersion data of the same variety of double-cropping rice in the second leaf and the fourth leaf in the second leaf tip period in multiple growth seasons are collected to form a sample first dispersion set and a sample second dispersion set. The sample first dispersion set and the sample second dispersion set can reflect the natural dispersion state of the leaf SPAD value of the double-cropping rice under normal growth conditions without pollution influence, and provide a reference basis for determining the preset first dispersion and the preset second dispersion. Then, statistical analysis is performed on the sample first dispersion set and the sample second dispersion set, such as calculating the average value or the median, to obtain standard values representing the SPAD value dispersion of the second leaf and the fourth leaf under normal growth conditions, which are respectively taken as the preset first dispersion and the preset second dispersion, representing the natural dispersion degree of the leaf SPAD value without pollution interference.

[0051] Then, the difference amplitude between the first dispersion measured at present and the preset first dispersion is calculated to obtain the first leaf pollution degree. Specifically, the pollution degree of the first leaf can be quantified by calculating the difference value or the ratio of the first dispersion and the preset first dispersion. When the first dispersion is significantly higher than the preset first dispersion, it indicates that the surface of the second leaf may be rough due to insufficient nitrogen fertilizer, which is easy to adsorb dust, thereby increasing the difference of the SPAD value at different positions, and the first leaf pollution degree is higher. Similarly, the difference amplitude between the second dispersion measured at present and the preset second dispersion is calculated to obtain the second leaf pollution degree. When the second dispersion deviates significantly from the preset second dispersion, it indicates that the surface state of the fourth leaf is abnormal, which may be polluted or affected by other factors. By quantifying the degree of deviation, the specific value of the second leaf pollution degree can be obtained.

[0052] By obtaining the first leaf pollution degree and the second leaf pollution degree, the nitrogen nutrition status of the double-cropping rice can be indirectly evaluated, which provides an important basis for subsequent fertilization decision-making.

[0053] Further, the first leaf image parameter array and the second leaf image parameter array are averaged and classified into leaf patterns to obtain a first leaf pattern array, and the first leaf contamination and the second leaf contamination are classified into leaf patterns to obtain a second leaf pattern array, including:

[0054] S310: The first leaf image parameter array and the second leaf image parameter array are respectively averaged and calculated to obtain a first leaf image parameter and a second leaf image parameter.

[0055] S320: A first leaf pattern classifier is constructed, wherein the first leaf pattern classifier includes a plurality of first leaf pattern classification branches, the input feature of each first leaf pattern classification branch is the ratio of the first leaf image parameter and the second leaf image parameter, and the output feature is the first leaf pattern.

[0056] S330: The ratio of the first leaf image parameter and the second leaf image parameter is calculated, the cycle category of the target double-season rice is combined, and the first leaf pattern classifier is input for integrated leaf pattern classification to obtain a first leaf pattern array, wherein the cycle category includes early rice or late rice.

[0057] S340: A second leaf pattern classifier is constructed, wherein the second leaf pattern classifier includes a plurality of second leaf pattern classification branches, the input feature of each second leaf pattern classification branch is the ratio of the first leaf contamination and the second leaf contamination, and the output feature is the second leaf pattern, and the number of the plurality of second leaf pattern classification branches is less than the number of the plurality of first leaf pattern classification branches.

[0058] S350: The ratio of the first leaf contamination and the second leaf contamination is calculated for integrated leaf pattern classification to obtain a second leaf pattern array.

[0059] In a preferred embodiment, the first leaf image parameter array and the second leaf image parameter array are respectively averaged and calculated to obtain a first leaf image parameter and a second leaf image parameter. Specifically, the SPAD values of the first leaf at a plurality of positions are summed and divided by the number of positions to obtain the first leaf image parameter; similarly, the SPAD values of the second leaf at a plurality of positions are summed and divided by the number of positions to obtain the second leaf image parameter. The two parameters respectively represent the average chlorophyll content level of the whole three-leaf and four-leaf, and can reflect the overall nitrogen nutrition status of the leaf.

[0060] Subsequently, a first leaf pattern classifier is constructed and used for leaf pattern classification. The first leaf pattern classifier comprises a plurality of first leaf pattern classification branches, each branch taking the ratio of the first leaf image parameters and the second leaf image parameters as input features and outputting a corresponding leaf pattern category. By calculating the ratio of the first leaf image parameters and the second leaf image parameters, the nitrogen allocation relationship between the third leaf and the fourth leaf can be reflected. When the ratio of the first leaf image parameters and the second leaf image parameters is greater than 1, it indicates that nitrogen is being transferred to the new leaf (the third leaf), and the plant may need additional fertilization; when the ratio of the first leaf image parameters and the second leaf image parameters is less than 1, it indicates that the SPAD value of the fourth leaf is higher than that of the third leaf, indicating that the nitrogen nutrition is sufficient or excessive, and the panicle fertilizer needs to be reduced or not applied. When performing leaf pattern classification, the cycle category (early rice or late rice) of the target double-cropping rice also needs to be combined, because there are differences in nitrogen demand of double-cropping rice in different seasons under the same SPAD ratio. The leaf pattern judgment criteria of early rice and late rice are also different due to differences in growth environment and growth characteristics. The calculated ratio is input into the first leaf pattern classifier, and the first leaf pattern array is obtained by integrating the results of a plurality of first leaf pattern classification branches.

[0061] At the same time, a second leaf pattern classifier is constructed and used for auxiliary leaf pattern classification. The second leaf pattern classifier also comprises a plurality of second leaf pattern classification branches, but the number of branches is less than that of the first leaf pattern classifier, reflecting its auxiliary nature. Each branch takes the ratio of the first leaf contamination and the second leaf contamination as input features and outputs a corresponding leaf pattern category. By calculating the ratio of the first leaf contamination and the second leaf contamination, the surface characteristic difference between the third leaf and the fourth leaf caused by the difference in nitrogen status can be reflected. When the ratio of the first leaf contamination and the second leaf contamination is greater than 1, it indicates that the contamination of the third leaf (the first leaf) is higher than that of the fourth leaf (the second leaf), indicating that the third leaf is more prone to adsorbing dust, the surface is rougher, and the nitrogen is insufficient, while the fourth leaf is relatively less prone to adsorbing dust and the nitrogen is sufficient, and the nitrogen is mainly concentrated in the old leaves and less transferred to the new leaves, and the panicle fertilizer needs to be reduced or not applied; when the ratio is less than 1, it indicates that the contamination of the third leaf is lower than that of the fourth leaf, indicating that the third leaf is not prone to adsorbing dust and the nitrogen nutrition status is good, and the nitrogen is being transferred from the old leaves (the fourth leaf) to the new leaves (the third leaf), and additional fertilization is needed to meet the growth needs. The calculated contamination ratio is input into the second leaf pattern classifier, and the second leaf pattern array is obtained by integrating the results of a plurality of second leaf pattern classification branches.

[0062] Through two different dimensions of leaf pattern classification methods, the nitrogen nutrition status of double-cropping rice can be comprehensively evaluated from two angles of SPAD value distribution and leaf surface state, providing a comprehensive and accurate basis for subsequent fertilization decision-making.

[0063] Further, a first leaf pattern classifier is constructed, comprising:

[0064] S321: According to the historical breeding data of the same family of double-cropping rice, a sample image parameter ratio set, a sample period category set, and a double-cropping rice leaf pattern under each sample image parameter ratio and sample period category are collected, and a sample first leaf pattern set is obtained by labeling, wherein the sample image parameter ratio is the ratio of the first leaf image parameter and the second leaf image parameter of the same family of double-cropping rice;

[0065] S322: The sample image parameter ratio set, the sample period category set, and the sample first leaf pattern set are divided to obtain multiple first mode classification training data, wherein there is an intersection between each two first mode classification training data;

[0066] S323: A plurality of first leaf pattern classification branches are constructed by machine learning;

[0067] S324: The multiple first leaf pattern classification branches are trained, verified, and tested respectively by using the multiple first mode classification training data to obtain a first leaf pattern classifier.

[0068] In a preferred embodiment, first, according to the historical breeding data of the same family of double-cropping rice, a sample image parameter ratio set and corresponding leaf pattern labeling are collected. Specifically, the first leaf image parameter and the second leaf image parameter of the same variety of double-cropping rice at the tip of the second leaf in the historical growing season are collected, the ratio of the first leaf image parameter and the second leaf image parameter is calculated to form a sample image parameter ratio set. At the same time, the period category (early rice or late rice) corresponding to each sample is collected to form a sample period category set. Then, the actual leaf growth state corresponding to each sample ratio is recorded as a leaf pattern label to obtain a sample first leaf pattern set. These historical data contain the growth performance of double-cropping rice under different nitrogen allocation states, providing basic data support for training the first leaf pattern classifier.

[0069] Then, the sample image parameter ratio set, the sample period category set, and the sample first leaf pattern set are divided to obtain multiple first mode classification training data. By dividing the sample data into multiple subsets, ensuring that there is a certain intersection between each two subsets, the generalization ability of the model can be enhanced, overfitting can be avoided, and the limited historical data can be fully utilized. Then, machine learning techniques are used to construct multiple first leaf pattern classification branches. Different first leaf pattern classification branches can be different types of classification algorithms, such as decision trees, support vector machines, neural networks, etc., or different parameter configurations of the same algorithm. The purpose of constructing multiple first leaf pattern classification branches is to form an ensemble learning model, which improves the overall classification performance by combining the prediction results of multiple base classifiers.

[0070] After that, multiple first leaf pattern classification branches are trained, validated and tested using multiple copies of the first mode classification training data. Each classification branch is trained using different training data, and then the parameters are adjusted through the validation set, and finally the performance is evaluated on the test set. After this series of processes, the first leaf pattern classifier capable of accurately identifying different leaf patterns is finally obtained. The first leaf pattern classifier can predict the current double-cropping rice leaf pattern according to the ratio of the first leaf image parameters and the second leaf image parameters, and provide a scientific basis for fertilization decision-making.

[0071] Similar to the construction of the first leaf pattern classifier, the second leaf pattern classifier also uses the same construction method. First, according to the historical cultivation data of the same family double-cropping rice, the ratio of the first leaf pollution degree and the second leaf pollution degree is collected to form a sample pollution degree ratio set; At the same time, the period category (early rice or late rice) corresponding to each sample is collected to form a sample period category set, and the actual leaf state under each pollution degree ratio and period category combination is recorded as a leaf pattern label, and a sample second leaf pattern set is obtained. Then, the sample pollution degree ratio set, the sample period category set and the sample second leaf pattern set are divided to ensure that there is an intersection between the training data subsets. Then, multiple second leaf pattern classification branches are constructed using machine learning techniques, which can use different types of classification algorithms or different parameter configurations of the same algorithm. After that, the training data is used to train, validate and test each classification branch, and finally a second leaf pattern classifier is formed that can predict the leaf pattern according to the ratio of the first leaf pollution degree and the second leaf pollution degree and the period category.

[0072] In this way, by combining the results of the first leaf pattern classifier and the second leaf pattern classifier, the nitrogen nutrition status of double-cropping rice can be comprehensively evaluated from two dimensions of SPAD value and leaf surface state, providing a reliable basis for the decision of panicle fertilization.

[0073] Further, according to the first leaf pattern array and the second leaf pattern array, a fused leaf pattern is obtained by processing, and the double-cropping rice panicle fertilization is performed, including:

[0074] S410: In the first leaf pattern array and the second leaf pattern array, select the leaf pattern with the largest occurrence ratio to obtain a fused leaf pattern;

[0075] S420: According to the fused leaf pattern, the double-cropping rice panicle fertilization is performed.

[0076] In a preferred embodiment, first, the leaf pattern with the highest occurrence frequency is selected from the first leaf pattern array and the second leaf pattern array to obtain a fused leaf pattern. Specifically, the various leaf patterns in the first leaf pattern array are counted, and the leaf pattern with the highest frequency is selected as the judgment result of the first dimension; similarly, the various leaf patterns in the second leaf pattern array are counted, and the leaf pattern with the highest frequency is selected as the judgment result of the second dimension. Then, the leaf patterns of the two dimensions are fused to form the final fused leaf pattern. This selection method based on the occurrence frequency can reduce the misjudgment risk of a single classifier through the integration of multiple classifiers, and improve the accuracy and stability of the leaf pattern classification.

[0077] Subsequently, according to the obtained fused leaf pattern, the application strategy of the panicle fertilizer of double-cropping rice is determined. Different fused leaf patterns correspond to different fertilization decisions, for example, when the fused leaf pattern shows that the leaf color of the third leaf from the bottom is similar to that of the fourth leaf from the bottom, the panicle fertilizer is normally applied according to the original calculation value; when the fused leaf pattern shows that the leaf color of the third leaf from the bottom is lighter than that of the fourth leaf from the bottom, the application amount of the panicle fertilizer is reduced; when the fused leaf pattern shows that the leaf color of the third leaf from the bottom is light, no panicle fertilizer is applied; when the fused leaf pattern shows that the leaf color of the third leaf from the bottom is darker than that of the fourth leaf from the bottom, the application amount of the panicle fertilizer is appropriately increased; and when the fused leaf pattern shows that the leaf color of the third leaf from the bottom is dark, the application amount of the panicle fertilizer is significantly increased.

[0078] Through the precise fertilization method based on the leaf pattern, the panicle fertilizer of double-cropping rice can be controlled according to the current actual nitrogen nutrition status, avoiding the problems of fertilizer waste or deficiency caused by blind experience fertilization, and improving the precision of the panicle fertilizer application of double-cropping rice.

[0079] As shown in Example 2, Figure 2 based on the same inventive concept of the leaf image pattern classification method for panicle fertilization of double-cropping rice provided in Example 1, the present application also provides a leaf image pattern classification system for panicle fertilization of double-cropping rice, comprising:

[0080] The leaf collecting module 11 is used for collecting the first leaf image parameter array and the second leaf image parameter array at a plurality of positions of the first leaf and the second leaf in a preset time window of the target double-cropping rice.

[0081] The pollution degree analysis module 12 is used for performing image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discrete degree and a second discrete degree, and combining a preset first discrete degree and a preset second discrete degree to calculate a first leaf pollution degree and a second leaf pollution degree.

[0082] The mode classification module 13 is configured to average the first leaf image parameter array and the second leaf image parameter array, and classify leaf modes to obtain a first leaf mode array; and classify leaf modes according to the first leaf contamination degree and the second leaf contamination degree to obtain a second leaf mode array.

[0083] The fertilization decision module 14 is configured to process the leaf modes according to the first leaf mode array and the second leaf mode array to obtain a leaf mode, and make a decision on application of a panicle fertilizer for double-cropping rice.

[0084] Further, the leaf collection module 11 includes the following execution steps:

[0085] At the second leaf and the third leaf of the target double-cropping rice, a plurality of positions of the first leaf and the second leaf are selected, wherein the first leaf is the third leaf, and the second leaf is the fourth leaf.

[0086] At the plurality of positions of the first leaf and the second leaf, SPAD parameters are respectively tested and collected to obtain a first leaf image parameter array and a second leaf image parameter array.

[0087] Further, the contamination degree analysis module 12 includes the following execution steps:

[0088] The first leaf image parameter array is subjected to image parameter dispersion analysis calculation to obtain a first dispersion degree, wherein the image parameter dispersion analysis is performed by calculating the standard deviation of a plurality of leaf image parameters in the first leaf image parameter array.

[0089] The second leaf image parameter array is subjected to image parameter dispersion analysis calculation to obtain a second dispersion degree.

[0090] Further, the contamination degree analysis module 12 further includes the following execution steps:

[0091] The first dispersion degree and the second dispersion degree of the same family of double-cropping rice in a preset time window in a historical time are obtained to obtain a sample first dispersion degree set and a sample second dispersion degree set.

[0092] The preset first dispersion degree and the preset second dispersion degree are calculated according to the sample first dispersion degree set and the sample second dispersion degree set.

[0093] The first leaf contamination degree and the second leaf contamination degree are obtained by respectively calculating the amplitudes of the first dispersion degree and the second dispersion degree deviating from the preset first dispersion degree and the preset second dispersion degree.

[0094] Further, the mode classification module 13 includes the following execution steps:

[0095] The first leaf image parameter array and the second leaf image parameter array are respectively subjected to average calculation processing to obtain a first leaf image parameter and a second leaf image parameter.

[0096] constructing a first leaf pattern classifier, wherein the first leaf pattern classifier comprises a plurality of first leaf pattern classification branches, an input feature of each first leaf pattern classification branch is a ratio of the first leaf image parameter and the second leaf image parameter, and an output feature is the first leaf pattern;

[0097] calculating the ratio of the first leaf image parameter and the second leaf image parameter, combining the cycle category of the target double-cropping rice, inputting the first leaf pattern classifier for integrated leaf pattern classification, and obtaining a first leaf pattern array, wherein the cycle category includes early rice or late rice;

[0098] constructing a second leaf pattern classifier, wherein the second leaf pattern classifier comprises a plurality of second leaf pattern classification branches, an input feature of each second leaf pattern classification branch is a ratio of the first leaf contamination degree and the second leaf contamination degree, and an output feature is the second leaf pattern, and the number of the plurality of second leaf pattern classification branches is less than the number of the plurality of first leaf pattern classification branches;

[0099] calculating the ratio of the first leaf contamination degree and the second leaf contamination degree, and performing integrated leaf pattern classification to obtain a second leaf pattern array.

[0100] Further, the pattern classification module 13 further includes the following execution steps:

[0101] According to historical cultivation data of the same family double-cropping rice, a sample image parameter ratio set, a sample cycle category set, and a leaf pattern of double-cropping rice under each sample image parameter ratio and sample cycle category are collected, and a sample first leaf pattern set is labeled and obtained, wherein the sample image parameter ratio is a ratio of the first leaf image parameter and the second leaf image parameter of the same family double-cropping rice;

[0102] The sample image parameter ratio set, the sample cycle category set, and the sample first leaf pattern set are divided to obtain a plurality of first pattern classification training data, wherein there is an intersection between every two first pattern classification training data;

[0103] A plurality of first leaf pattern classification branches are constructed by using machine learning;

[0104] The plurality of first leaf pattern classification branches are trained, verified, and tested by using the plurality of first pattern classification training data respectively, and a first leaf pattern classifier is obtained.

[0105] Further, the fertilization decision module 14 includes the following execution steps:

[0106] In the first leaf pattern array and the second leaf pattern array, a leaf pattern with the largest occurrence ratio is selected to obtain a fused leaf pattern.

[0107] According to the fusion leaf mode, a panicle fertilizer is applied for double-cropping rice.

[0108] It should be noted that in the above examples, the description of each example has its own focus, and the parts not described in detail in a certain example can be referred to the related description of other examples.

[0109] Those skilled in the art will understand that embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) containing computer usable program code.

[0110] The present application is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce an apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by the flow or flows and / or block or blocks.

[0111] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by the flow or flows and / or block or blocks.

[0112] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by the flow or flows and / or block or blocks.

[0113] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application.

[0114] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims, the application can be practiced otherwise than as specifically described herein.

Claims

1. A leaf image pattern classification method for double-season rice ear fertilization, characterized in that: The method comprises: In a preset time window of the target double-cropping rice, collecting and obtaining a first leaf image parameter array and a second leaf image parameter array at multiple positions of the first leaf and the second leaf; Performing image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discreteness and a second discreteness, and calculating a first leaf contamination degree and a second leaf contamination degree by combining a preset first discreteness and a preset second discreteness; The method further comprises: performing averaging processing and leaf pattern classification on the first leaf image parameter array and the second leaf image parameter array to obtain a first leaf pattern array; and performing leaf pattern classification according to the first leaf contamination degree and the second leaf contamination degree to obtain a second leaf pattern array, comprising: Performing averaging calculation processing on the first leaf image parameter array and the second leaf image parameter array respectively to obtain first leaf image parameters and second leaf image parameters; Constructing a first leaf pattern classifier, wherein the first leaf pattern classifier includes a plurality of first leaf pattern classification branches, an input feature of each first leaf pattern classification branch is a ratio of a first leaf image parameter to a second leaf image parameter, and an output feature is a first leaf pattern; calculating a ratio of the first leaf image parameter to the second leaf image parameter, combining the cycle category of the target double-season rice, and inputting the ratio into the first leaf pattern classifier for integrated leaf pattern classification to obtain a first leaf pattern array, wherein the cycle category includes early rice or late rice; Constructing a second leaf pattern classifier, wherein the second leaf pattern classifier includes a plurality of second leaf pattern classification branches, an input feature of each second leaf pattern classification branch is a ratio of the first leaf contamination degree to the second leaf contamination degree, an output feature is a second leaf pattern, and the number of the plurality of second leaf pattern classification branches is less than the number of the plurality of first leaf pattern classification branches; calculating a ratio of the first leaf contamination degree to the second leaf contamination degree, performing integrated leaf pattern classification, and obtaining a second leaf pattern array; According to the first leaf pattern array and the second leaf pattern array, leaf patterns are obtained by processing and double-season rice ear fertilizer is applied.

2. The leaf image pattern classification method for double-season rice panicle fertilization according to claim 1, characterized in that: In a preset time window of the target double-cropping rice, the first leaf image parameter array and the second leaf image parameter array are acquired at multiple positions of the first leaf and the second leaf, including: At the stage when the second leaf of the target double-cropping rice is about to emerge, selecting multiple positions of the first leaf and the second leaf of the target double-cropping rice, wherein the first leaf is the third leaf and the second leaf is the fourth leaf; SPAD parameters are tested and collected at multiple positions of the first leaf and the second leaf respectively to obtain a first leaf image parameter array and a second leaf image parameter array.

3. The leaf image pattern classification method for double-season rice panicle fertilization according to claim 1, characterized in that: Performing image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discreteness and a second discreteness includes: performing an image parameter discrete analysis calculation on the first leaf image parameter array to obtain a first discreteness, wherein the image parameter discrete analysis is performed by calculating a standard deviation of a plurality of leaf image parameters in the first leaf image parameter array; Perform image parameter discrete analysis calculation on the second leaf image parameter array to obtain a second discrete degree.

4. The leaf image pattern classification method for double-season rice panicle fertilization according to claim 1, characterized in that: Calculating the first leaf contamination degree and the second leaf contamination degree by combining the preset first discreteness and the preset second discreteness includes: Obtain the first discreteness and the second discreteness of the same family of double-season rice in a preset time window within a historical period, and obtain a sample first discreteness set and a sample second discreteness set; Calculate and configure a preset first discreteness and a preset second discreteness according to the sample first discreteness set and the sample second discreteness set; The magnitudes by which the first dispersion and the second dispersion deviate from the preset first dispersion and the preset second dispersion are calculated respectively to obtain a first leaf contamination degree and a second leaf contamination degree.

5. The leaf image pattern classification method for double-season rice panicle fertilization according to claim 1, characterized in that: Construct the first leaf pattern classifier, including: Based on historical cultivation data of double-season rice of the same family, a set of sample image parameter ratios and a set of sample period categories are collected. The leaf patterns of double-season rice under each sample image parameter ratio and sample period category are collected and annotated to obtain a set of sample first leaf patterns, wherein the sample image parameter ratio is the ratio of the first leaf image parameter to the second leaf image parameter of the double-season rice of the same family. Dividing the sample image parameter ratio set, the sample period category set, and the sample first leaf pattern set to obtain a plurality of first pattern classification training data, wherein there is an intersection between every two first pattern classification training data; Using machine learning, multiple first leaf pattern classification branches are constructed; The plurality of first pattern classification training data are respectively used to train, verify and test the plurality of first leaf pattern classification branches to obtain a first leaf pattern classifier.

6. The leaf image pattern classification method for double-season rice panicle fertilization according to claim 1, characterized in that: Processing the first leaf pattern array and the second leaf pattern array to obtain a fused leaf pattern and applying double-season rice ear fertilizer includes: In the first blade pattern array and the second blade pattern array, selecting the blade pattern with the largest number of occurrences to obtain a fused blade pattern; According to the fused leaf pattern, double-season rice panicle fertilizer is applied.

7. A leaf image pattern classification system for double-season rice ear fertilization, characterized in that: A system for implementing the leaf image pattern classification method for double-season rice panicle fertilization according to any one of claims 1 to 6, comprising: A leaf acquisition module is used to acquire a first leaf image parameter array and a second leaf image parameter array at multiple positions of the first leaf and the second leaf during a preset time window of the target double-cropping rice; a pollution degree analysis module, configured to perform image parameter discrete analysis on the first leaf image parameter array and the second leaf image parameter array to obtain a first discrete degree and a second discrete degree, and calculate the first leaf pollution degree and the second leaf pollution degree by combining a preset first discrete degree and a preset second discrete degree; A leaf pattern classification module is configured to perform averaging processing and leaf pattern classification on the first leaf image parameter array and the second leaf image parameter array to obtain a first leaf pattern array, and perform leaf pattern classification based on the first leaf contamination degree and the second leaf contamination degree to obtain a second leaf pattern array, including: Performing averaging calculation processing on the first leaf image parameter array and the second leaf image parameter array respectively to obtain first leaf image parameters and second leaf image parameters; Constructing a first leaf pattern classifier, wherein the first leaf pattern classifier includes a plurality of first leaf pattern classification branches, an input feature of each first leaf pattern classification branch is a ratio of a first leaf image parameter to a second leaf image parameter, and an output feature is a first leaf pattern; calculating a ratio of the first leaf image parameter to the second leaf image parameter, combining the cycle category of the target double-season rice, and inputting the ratio into the first leaf pattern classifier for integrated leaf pattern classification to obtain a first leaf pattern array, wherein the cycle category includes early rice or late rice; Constructing a second leaf pattern classifier, wherein the second leaf pattern classifier includes a plurality of second leaf pattern classification branches, an input feature of each second leaf pattern classification branch is a ratio of the first leaf contamination degree to the second leaf contamination degree, an output feature is a second leaf pattern, and the number of the plurality of second leaf pattern classification branches is less than the number of the plurality of first leaf pattern classification branches; calculating a ratio of the first leaf contamination degree to the second leaf contamination degree, performing integrated leaf pattern classification, and obtaining a second leaf pattern array; The fertilization decision module is used to process and obtain leaf patterns according to the first leaf pattern array and the second leaf pattern array, and apply double-season rice ear fertilizer.

Images