Unmanned aerial vehicle disease and pest monitoring system based on big data

Abstract

The invention discloses an unmanned aerial vehicle disease and pest monitoring system based on big data, and belongs to the technical field of disease and pest monitoring, and the system comprises an unmanned aerial vehicle path flight module which is used for determining flight routes and ensuring that each flight route can fully cover vegetation of a corresponding sub-monitoring area; the unmanned aerial vehicle acquisition module is used for acquiring associated parameter information of a monitoring area; the analysis processing module is used for processing and analyzing the obtained associated parameter information to calculate an abnormal value, and judging whether plant diseases and insect pests exist in the vegetation according to the abnormal value condition; and the evaluation module is used for evaluating the hazard risk of plant diseases and insect pests of the vegetation according to the judgment result. The plant diseases and insect pests are automatically monitored through the unmanned aerial vehicle technology, manual detection is not needed, the subjective influence can be eliminated, the detection result is more real, comprehensive analysis is conducted through the obtained image information, spectral information, thermal imaging information and the like of the monitoring area, the monitoring area can be detected in an omnibearing mode, and the detection efficiency is improved. And the detection accuracy is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of pest and disease monitoring, and particularly relates to a drone pest and disease monitoring system based on big data. Background Art

[0002] Pests and diseases of crops seriously affect the yield and quality of agricultural production and pose a continuous threat to global food security.

[0003] When monitoring pests and diseases, most of the existing methods are manual investigations. The efficiency of manual investigations is low, and the investigation results are greatly affected by subjective experience. Different personnel may have different judgment criteria for pests and diseases, resulting in difficulties in ensuring the accuracy and consistency of monitoring results. There are also some fixed monitoring points set up in farmland, such as equipped with trapping devices, sensors, etc. to obtain data such as pest species and environment, so as to monitor pests and diseases. This method has a limited monitoring range and can only reflect the pest and disease conditions in a small area around the monitoring point. For large areas of farmland, when the monitoring points are sparsely distributed, it is difficult to comprehensively grasp the spatial distribution of pests and diseases. Summary of the Invention

[0004] The purpose of the present invention is to provide a drone pest and disease monitoring system based on big data to solve the problems faced in the above background art.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A drone pest and disease monitoring system based on big data, the monitoring system includes:

[0007] A drone path flight module, which determines multiple flight routes according to the shape and size of the monitoring area to ensure that each flight route can fully cover the vegetation in the corresponding sub-monitoring area;

[0008] A drone acquisition module, which uses a drone to carry a variety of sensors to obtain the associated parameter information of the monitoring area;

[0009] An analysis and processing module, which is used to process and analyze the obtained associated parameter information, thereby calculating the abnormal value, and judging whether there are pests and diseases on the vegetation according to the abnormal value situation;

[0010] An evaluation module, which judges the hazard risk of the pests and diseases on the vegetation when it is judged that there are pests and diseases.

[0011] Further, the associated parameter information includes image information, spectral information, and thermal imaging information.

[0012] Further, the working method of the analysis and processing module is:

[0013] Obtain the vegetation physiological parameter value R based on the acquired spectral information S , obtain the vegetation temperature anomaly parameter value R based on the acquired thermal imaging information T , obtain the characteristic parameter value R through the acquired image information C and the pest and disease parameter value R G ;

[0014] Obtain the anomaly value R through the formula ;

[0015] When R > R m , it is determined that there are pests and diseases in the plants, where γ1, γ2, and γ3 are preset coefficients, ω is a conversion coefficient, and R m is a preset anomaly threshold.

[0016] Furthermore, the method for obtaining the vegetation physiological parameter value R S is as follows:

[0017] Based on the acquired spectral information, obtain the reflectance at different spectral bands, and thus calculate the normalized difference vegetation index S of the index NDVI and the chlorophyll content S Chi ;

[0018] Obtain the vegetation physiological parameter value R through the formula ; S ;

[0019] Among them, and are the standard normalized difference vegetation index and the standard chlorophyll content of the current stage index respectively.

[0020] Furthermore, the method for obtaining the vegetation temperature anomaly parameter value R T is as follows:

[0021] Based on the acquired thermal imaging information, obtain the temperature distribution map of the corresponding sub-monitoring area, and identify the area T of the abnormal temperature area in the temperature distribution map s , and obtain the temperature anomaly parameter value R through the formula ; T ;

[0022] Among them, s0 is the total area of the temperature distribution map, w1 and w2 are proportionality coefficients, T v is the average temperature of the temperature distribution map, is the preset standard temperature of the current stage, and ΔT is the temperature reference value.

[0023] Furthermore, the characteristic parameter value R C and the pest and disease parameter value R GThe acquisition method is as follows:

[0024] According to the acquired image information, extract the color features in the image through RGB technology, extract the texture features in the image through the gray-level co-occurrence matrix, and extract the contour of the vegetation in the image through the Canny edge detection technology, so as to obtain the shape features;

[0025] Based on the neural network model, train a state score model for color features, texture features, and shape features, and input the acquired color features, texture features, and shape features into the corresponding state score models to obtain the corresponding color state score C colour , texture state score C texture and shape state score C shape , and through the formula R C =C colour +C texture +C shape to obtain the feature parameter value R C ;

[0026] At the same time, based on the image recognition technology, identify the type n of pests in the image and the number D of each type of pest during each collection i , and thus through the formula to obtain the pest and disease parameter value R G ;

[0027] where ρ is the error coefficient, ∈ i is the hazard risk coefficient of the i-th type of pest to the vegetation, which is obtained according to the relevant pest data in the big data, and i∈(1, n).

[0028] Furthermore, the working method of the evaluation module is as follows:

[0029] When it is judged that there are pests and diseases, obtain the abnormal value R of the previous n collections, and thus formulate the curve function R(x) of the abnormal value changing with the number of times;

[0030] Through the formula to obtain the judgment value K,

[0031] Compare the obtained judgment value K with the preset judgment thresholds K v , K b :

[0032] When K∈(0, K v ), it is judged that the hazard risk of the pests and diseases of the vegetation is relatively light, and it is marked as the first-level risk level;

[0033] When K∈[K v , K b , it is judged that the hazard risk of the pests and diseases of the vegetation is relatively serious, and it is marked as the second-level risk level;

[0034] When \(K\in(K b , +\infty)\), it is determined that the risk of damage to the vegetation by pests and diseases is very serious, and it is marked as the third-level risk level;

[0035] Among them, \(\max R'(x)\) is the maximum slope of \(R(x)\), \(x_1\) is the first collection, \(x_2\) is the last collection, is the average height of the vegetation at the last collection, is the average height of the vegetation at the first collection, is the value of the pest and disease parameter at the last collection, is the value of the pest and disease parameter at the first collection.

[0036] Furthermore, the evaluation module is also used to judge the growth stability of the vegetation in each sub-monitoring area when it is determined that there are no pests and diseases in the monitoring area:

[0037] When there are no pests and diseases, obtain the outlier \(R\) of each sub-monitoring area, and through the formula obtain the stability coefficient \(\delta\);

[0038] When \(\delta\gt\delta_0\), it is determined that the growth stability of the vegetation in this sub-monitoring area is poor;

[0039] Among them, \(\mu_1\) and \(\mu_2\) are preset coefficients, \(M\) is the number of sub-monitoring areas, \(R j is the outlier of the \(j\)-th sub-monitoring area, and \(j\in(1, M)\), is the standard outlier of the \(j\)-th sub-monitoring area, and \(\delta_0\) is the preset stability coefficient threshold.

[0040] Advantages of the present invention:

[0041] The present invention automatically monitors the pests and diseases of the vegetation through drone technology, without manual detection, which can eliminate the subjective influence and make the detection result more real. At the same time, through comprehensive analysis of the obtained monitoring area image information, spectral information, thermal imaging information, etc., the monitoring area can be detected in all directions, improving the accuracy of detection.

[0042] The present invention can judge whether there are pests and diseases in the monitoring area according to the obtained outlier, and when there are pests and diseases, analyze according to the change of the outlier and the growth of the vegetation to judge the severity and trend of the pests and diseases, so as to timely deal with the pests and diseases to ensure the normal growth of the vegetation. When there are no pests and diseases, judge the growth stability of a single sub-monitoring area in the entire monitoring area according to the obtained outlier situation, so as to facilitate timely intervention in the area with poor growth stability and avoid or reduce the occurrence of subsequent abnormal vegetation growth phenomena.

[0043] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0045] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0047] In one embodiment, a UAV pest and disease monitoring system based on big data is disclosed. As Figure 1 shown, the monitoring system includes:

[0048] A UAV path flight module that determines multiple flight routes according to the shape and size of the monitoring area to ensure that each flight route can fully cover the vegetation in the corresponding sub-monitoring area;

[0049] A UAV acquisition module that uses a UAV to carry a variety of sensors to obtain the associated parameter information of the monitoring area. The associated parameter information includes image information, spectral information, and thermal imaging information;

[0050] An analysis and processing module that is used to process and analyze the obtained associated parameter information, thereby calculating the abnormal values, and judging whether there are pests and diseases in the vegetation according to the abnormal value situation;

[0051] An evaluation module that judges the harm risk of pests and diseases in the vegetation when it is judged that there are pests and diseases. The evaluation module is also used to judge the growth stability of the vegetation in each sub-monitoring area when it is judged that there are no pests and diseases in the monitoring area.

[0052] Through the above technical solution, the present application automatically monitors the diseases and pests of vegetation through drone technology, eliminating the need for manual detection, excluding subjective influence, making the detection results more authentic. At the same time, through comprehensive analysis of the obtained image information, spectral information, thermal imaging information, etc. of the monitoring area, the monitoring area can be detected in all directions, improving the accuracy of detection. In addition, the present application can also judge whether there are diseases and pests in the monitoring area according to the obtained abnormal values, and analyze the severity and trend of the diseases and pests according to the abnormal value situation and the growth situation of the vegetation when there are diseases and pests, so as to timely carry out corresponding treatments on the diseases and pests to ensure the normal growth of the vegetation. When there are no diseases and pests, the growth stability of individual sub-monitoring areas in the entire monitoring area is judged according to the obtained abnormal value situation, so as to facilitate timely intervention in areas with poor growth stability and avoid or reduce the occurrence of subsequent abnormal vegetation growth phenomena.

[0053] The working method of the analysis and processing module is as follows: obtaining the vegetation physiological parameter value R according to the obtained spectral information S , obtaining the vegetation temperature anomaly parameter value R according to the obtained thermal imaging information T , obtaining the characteristic parameter value R through the obtained image information C and the disease and pest parameter value R G ;

[0054] Through the formula obtaining the abnormal value R;

[0055] When R > R m , it is determined that there are diseases and pests in the vegetation, where γ1, γ2, and γ3 are preset coefficients, ω is a conversion coefficient, and R m is a preset abnormal threshold;

[0056] The method for obtaining the vegetation physiological parameter value R S is as follows: according to the obtained spectral information, obtaining the reflectance under different spectral bands, and thus calculating the normalized difference vegetation index S NDVI and the chlorophyll content S Chi ;

[0057] Through the formula obtaining the vegetation physiological parameter value R S ;

[0058] Among them, and are the standard normalized difference vegetation index and the standard chlorophyll content of the current stage index respectively;

[0059] The vegetation temperature anomaly parameter value R TThe acquisition method is as follows: according to the acquired thermal imaging information, obtain the temperature distribution map of the corresponding sub-monitoring area, and identify the area T of the abnormal temperature area in the temperature distribution map. s , through the formula obtain the temperature anomaly parameter value R T ;

[0060] where s0 is the total area of the temperature distribution map, w1 and w2 are proportionality coefficients, T v is the average temperature of the temperature distribution map, is the preset standard temperature at the current stage, and ΔT is the temperature reference value

[0061] The characteristic parameter value R C and the pest damage parameter value R G The acquisition method is as follows: according to the acquired image information, extract the color features in the image through RGB technology, extract the texture features in the image through the gray-level co-occurrence matrix, and extract the contour of the vegetation in the image through the Canny edge detection technology, so as to obtain the shape features;

[0062] Based on the neural network model, train the state score model for color features, texture features and shape features, and input the obtained color features, texture features and shape features into the corresponding state score models to obtain the corresponding color state score C colour , texture state score C texture and shape state score C shape , through the formula R C = C colour + C texture + C shape obtain the characteristic parameter value R C ;

[0063] At the same time, based on the image recognition technology, identify the types n of pests in the image and the number D of each type of pest during each collection i , thus through the formula obtain the pest damage parameter value R G ;

[0064] where ρ is the error coefficient, ∈ i is the hazard risk coefficient of the i-th type of pest to the vegetation, which is obtained according to the relevant pest data in the big data, and i ∈ (1, n).

[0065] Through the above technical solution, this embodiment mainly provides a method for obtaining outliers. Generally, when vegetation is affected by pests and diseases, its photosynthesis is affected, resulting in a significant decrease in chlorophyll content and NDVI value compared to the healthy area. Similarly, when there are pests and diseases, due to abnormal physiological metabolism, the temperature of the diseased or pest-infested part will differ from that of the healthy part. Moreover, when affected by pests and diseases, due to abnormal growth, the vegetation leaves may show curling, wrinkling, yellowing and other phenomena. Therefore, after the drone obtains image information, spectral information and thermal imaging information, according to the obtained spectral information, the reflectance under different spectral bands is obtained, and then the normalized difference vegetation index S of the index is calculated. NDVI and chlorophyll content S Chi , and then through the formula the vegetation physiological parameter value R is obtained. S In the formula, and are the standard normalized difference vegetation index and standard chlorophyll content of the current stage index respectively, which can be obtained according to the relevant vegetation data of normal growth in the current stage in the big data. It can be seen from the formula that when the vegetation physiological parameter value R S is smaller, it indicates a greater possibility of pest infestation; similarly, according to the obtained thermal imaging information, the temperature distribution map of the corresponding sub-monitoring area is obtained, and the area T of the abnormal temperature area in the temperature distribution map is identified. s , through the formula the temperature anomaly parameter value R is obtained. T In the formula, s0 is the total area of the temperature distribution map, w1 and w2 are proportionality coefficients, which are obtained according to empirical data, and the sum of the two is equal to one. For example, w1 = 0.45 and w2 = 0.55 can be set. T v is the average temperature of the temperature distribution map, is the preset standard temperature of the current stage, and ΔT is the temperature reference value. Both can be obtained according to the relevant vegetation data of normal growth in the current stage in the big data. It can be known that when the temperature anomaly parameter value R T is larger, it indicates a greater possibility of pest infestation; similarly, according to the obtained image information, the color features in the image are extracted through RGB technology, the texture features in the image are extracted through the gray-level co-occurrence matrix, and the contour of the vegetation in the image is extracted through the Canny edge detection technology, so as to obtain the shape features; based on the neural network model, a state score model for color features, texture features and shape features is trained, and the obtained color features, texture features and shape features are input into the corresponding state score models to obtain the corresponding color state score C colour , texture state score C texture and shape state score C shape . Finally, through the formula R C =C colour +Ctexture +C shape Obtain the characteristic parameter value R C , when the status score is smaller, it indicates that the vegetation status is worse. Therefore, when the characteristic parameter value R C is smaller, it indicates that the possibility of pest infestation is greater; finally, based on the image recognition technology, identify the type n of pests in the image during each collection and the number D of each type of pest i , and thus through the formula Obtain the pest and disease parameter value R G , where ρ is the error coefficient, which is adaptively determined according to the weather conditions of the day. For example, if the weather is sunny and windless on the day, ρ can be set to 1. If the weather is sunny but windy on the day, the detection error may be large, then ρ = 1.2, ∈ i is the hazard risk coefficient of the i-th type of pest to the vegetation, which is obtained based on the relevant pest data in the big data. It can be known that when the pest and disease parameter value R G is larger, the possibility of pest infestation in the vegetation is greater; finally, through Obtain the outlier R, where ω is the conversion coefficient, which converts the pest and disease parameter value R G into a numerical value for easy calculation. The preset coefficients γ1, γ2, and γ3 can be determined according to experience. It can be known that the larger the outlier R, the greater the possibility of pest infestation in the vegetation. Therefore, compare it with the outlier threshold R m preset according to the empirical data m . When R > R

[0066] , it is determined that there are pests and diseases in the vegetation. In this way, it is possible to comprehensively analyze the physiological condition, growth condition, temperature anomaly condition of the vegetation, as well as the types and quantities of corresponding pests and diseases, and more accurately determine whether there is a pest situation in the monitoring area

[0067] The working method of the evaluation module is: when it is determined that there are pests and diseases, obtain the outlier R of the previous n collections, and thus formulate the outlier change curve function R(x) with the number of times Obtain the judgment value K through the formula

[0068] Compare the obtained judgment value K with the preset judgment thresholds K v , K b :

[0069] When K ∈ (0, K v ), it is determined that the hazard risk of the pests and diseases of the vegetation is relatively light, and it is marked as the first-level risk level

[0070] When K ∈ [K v , K b , it is determined that the hazard risk of the pests and diseases of the vegetation is relatively serious, and it is marked as the second-level risk level

[0071] When K ∈ (K b , +∞), it is determined that the risk of damage caused by plant diseases and insect pests to the vegetation is very serious, and it is marked as the third-level risk level;

[0072] Among them, maxR'(x) is the maximum slope of R(x), x1 is the first collection, x2 is the last collection, is the average height of the vegetation at the last collection, is the average height of the vegetation at the first collection, is the value of the plant disease and insect pest parameter at the last collection, is the value of the plant disease and insect pest parameter at the first collection;

[0073] When there are no plant diseases and insect pests, the abnormal value R of each sub-monitoring area is obtained, and the stability coefficient δ is obtained through the formula ;

[0074] When δ > δ0, it is determined that the growth stability of the vegetation in this sub-monitoring area is poor;

[0075] Among them, μ1 and μ2 are preset coefficients, M is the number of sub-monitoring areas, R j is the abnormal value of the j-th sub-monitoring area, and j ∈ (1, M), is the standard abnormal value of the j-th sub-monitoring area, and δ0 is the preset stability coefficient threshold.

[0076] Through the above technical solution, this embodiment provides a specific method for the evaluation module to work. First, when it is determined that there are plant diseases and insect pests, the risk level of the insect pests is determined. Specifically, the abnormal value R of the first n collections is obtained, and then the curve function R(x) of the abnormal value changing with the number of times is drawn up; through the formula the judgment value K is obtained. It can be seen from the formula that when the height difference of the vegetation at the last collection is smaller than that at the first collection, it means that it is more affected by plant diseases and insect pests. Similarly, when the cumulative value of the abnormal value is higher and the change trend is greater, it means that the plant diseases and insect pests are expanding. Similarly, when the value of the plant disease and insect pest parameter at the last collection has a significant increase compared with the first time, it means that the plant diseases and insect pests are further expanding. Therefore, the larger the judgment value K, the higher the risk of plant diseases and insect pests. Therefore, the obtained judgment value K is compared with the preset judgment thresholds K v , K b : When K ∈ (0, K v ), it is determined that the risk of damage caused by plant diseases and insect pests to the vegetation is relatively light, and it is marked as the first-level risk level; when K ∈ K v , K b , it is determined that the risk of damage caused by plant diseases and insect pests to the vegetation is relatively serious, and it is marked as the second-level risk level; when K ∈ (K b, +∞), it is determined that the risk of damage caused by plant diseases and insect pests to the vegetation is extremely serious and marked as the third-level risk level. By this means, when it is determined that there are plant diseases and insect pests, the severity and trend of the plant diseases and insect pests can be judged by analyzing the change of the abnormal value and the growth condition of the vegetation, so that the corresponding treatment can be carried out on the plant diseases and insect pests in time to ensure the normal growth of the vegetation. Then, when there are plant diseases and insect pests, the abnormal value R of each sub-monitoring area is obtained, and the stability coefficient δ is obtained through the formula ; according to the difference between the abnormal value of a single sub-area and the average value of the entire monitoring area, and at the same time, a comprehensive analysis is carried out according to the difference between the abnormal value of a single sub-area and the abnormal value of the current standard. Since the entire monitoring area is in the same environment, when the value of the stability coefficient is larger, it can be judged that the stability of a single sub-monitoring area relative to the entire monitoring area is poor. In this way, the growth stability of a single sub-monitoring area in the entire monitoring area can be judged, so as to facilitate timely intervention in the area with poor growth stability and avoid or reduce the occurrence of subsequent abnormal vegetation growth phenomena.

[0077] It should be noted that for the convenience of analysis and calculation, the calculations in the above formula are all calculations between the processed numbers. The number of acquisitions is determined according to the growth cycle of the vegetation, and the preset judgment threshold K v 、K b 、the standard abnormal value of each sub-monitoring area The preset stability coefficient threshold δ0 can all be determined according to historical data and empirical data.

[0078] The above content is only an example and explanation of the concept of the present invention. Those skilled in the art of this technology make various modifications or supplements to the described specific embodiments or use similar methods to replace them. As long as they do not deviate from the concept of the invention or exceed the scope defined by this claim book, they should all fall within the protection scope of the present invention.

Claims

1. A drone pest and disease monitoring system based on big data, characterized in that, The monitoring system includes: A drone path flight module, which determines multiple flight routes according to the shape and size of the monitoring area to ensure that each flight route can fully cover the vegetation in the corresponding sub-monitoring area; A drone acquisition module, which uses a drone to carry a variety of sensors to obtain the associated parameter information of the monitoring area; An analysis and processing module, which is used to process and analyze the obtained associated parameter information, calculate the abnormal value, and judge whether there are pests and diseases in the vegetation according to the abnormal value situation; An evaluation module, which judges the risk of damage caused by pests and diseases to the vegetation when it is judged that there are pests and diseases.

2. The drone pest and disease monitoring system based on big data according to claim 1, characterized in that, The associated parameter information includes image information, spectral information, and thermal imaging information.

3. The drone pest and disease monitoring system based on big data according to claim 2, wherein, The working method of the analysis and processing module is as follows: Derive the vegetation physiological parameter value R based on the acquired spectral information S , derive the vegetation temperature anomaly parameter value R based on the acquired thermal imaging information T , derive the characteristic parameter value R through the acquired image information C and the pest and disease parameter value R G ; Through the formula the outlier R is obtained; When R > R m it is determined that there are pests and diseases in the planting, where γ1, γ2, and γ3 are preset coefficients, ω is a conversion coefficient, and R m is a preset abnormal threshold.

4. A drone pest and disease monitoring system based on big data according to claim 3, characterized in that, The vegetation physiological parameter value R S The acquisition method is as follows: Based on the acquired spectral information, obtain the reflectance under different spectral bands, and thus calculate the normalized difference vegetation index S of the index NDVI and the chlorophyll content S Chi ; Derive the vegetation physiological parameter value R through the formula S ;​ Among them, and are the standard normalized difference vegetation index and the standard chlorophyll content of the current stage index, respectively.

5. The drone pest and disease monitoring system based on big data according to claim 3, characterized in that, The vegetation temperature anomaly parameter value R T The acquisition method is as follows: According to the acquired thermal imaging information, obtain the temperature distribution map of the corresponding sub-monitoring area, and identify the area T of the abnormal temperature area in the temperature distribution map s , through the formula obtain the temperature anomaly parameter value R T ; Among them, s0 is the total area of the temperature distribution map, w1 and w2 are proportionality coefficients, and T v is the average temperature of the temperature distribution map, is the standard temperature preset for the current stage, and ΔT is the temperature reference value.

6. The drone pest and disease monitoring system based on big data according to claim 3, wherein, The characteristic parameter value R C and the pest and disease parameter value R G The acquisition method is as follows: According to the obtained image information, the color features in the image are extracted by RGB technology, the texture features in the image are extracted by the gray-level co-occurrence matrix, and the contour of the vegetation in the image is extracted by the Canny edge detection technology, so as to obtain the shape features; Based on a neural network model, a state score model for color features, texture features, and shape features is trained. The obtained color features, texture features, and shape features are input into the corresponding state score models to obtain the corresponding color state score C colour , texture state score C texture and shape state score C shape . Through the formula R C = C colour + C texture + C shape , the feature parameter value R C is obtained; Meanwhile, based on the image recognition technology, identify the types \(n\) of pests in the image during each collection and the number \(D\) of each type of pest i , so as to obtain the pest and disease parameter value \(R\) through the formula ; G ; where ρ is the error coefficient, ∈ i is the hazard risk coefficient of the i-th pest to the vegetation, which is obtained based on the relevant pest data in the big data, and i ∈ (1, n).

7. The drone pest and disease monitoring system based on big data according to claim 3, wherein The working method of the evaluation module is as follows: When it is judged that there are pests and diseases, the abnormal values R collected in the previous n times are obtained, and the curve function R(x) of the change of the abnormal value with the number of times is formulated; Through the formula the judgment value K is obtained. Compare the obtained judgment value K with the preset judgment thresholds K v and K b as follows: When K ∈ (0, K v ), it is determined that the risk of damage to the vegetation by pests and diseases is relatively low, and it is marked as the first-level risk level; When K ∈ [K v , K b , it is determined that the risk of damage to the vegetation by pests and diseases is relatively serious and marked as the second-level risk level; When K ∈ (K b , +∞), it is determined that the risk of damage to vegetation by pests and diseases is extremely serious, and it is marked as the third-level risk level; Among them, maxR'(x) is the maximum slope of R(x), x1 is the first collection, and x2 is the last collection. is the average height of the vegetation at the last collection. is the average height of the vegetation at the first collection. is the value of the pest and disease parameter at the last collection. is the value of the pest and disease parameter at the first collection.

8. The drone pest and disease monitoring system based on big data according to claim 3, wherein, The evaluation module is also used to judge the growth stability of the vegetation in each sub-monitoring area when it is judged that there are no pests and diseases in the monitoring area: When there are no pests and diseases, obtain the anomaly value R of each sub-monitoring area, and through the formula obtain the stability coefficient δ; When δ > δ0, it is judged that the growth stability of the vegetation in this sub-monitoring area is poor; where μ1 and μ2 are preset coefficients, M is the number of sub-monitoring regions, R j is the outlier of the j-th sub-monitoring region, and j ∈ (1, M), is the standard outlier of the j-th sub-monitoring region, and δ0 is the preset stability coefficient threshold.

Images