A method for identifying diseases and insect pests of navel orange

Abstract

The invention relates to the technical field of remote sensing, and proposes a method for identifying navel orange pests and diseases, comprising the following steps: obtaining multispectral satellite image information of a first satellite, a second satellite and a third satellite within a critical growth period of the navel orange, and preprocessing the multispectral satellite image information; presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite and the third satellite, so as to obtain a multispectral satellite synthetic image of each time window within the critical growth period of the navel orange; performing vegetation index calculation on the multispectral satellite synthetic image of each time window, so as to obtain a time series vegetation index, and a large-scale navel orange planting area can be identified through the multispectral image time series, and on-board processing is performed in combination with a high-resolution satellite, so as to improve the identification efficiency and accuracy of navel orange pests and diseases.

Description

Technical Field

[0001] The invention relates to the technical field of satellite remote sensing, and in particular to a method for identifying diseases and insect pests of navel orange. Background Art

[0002] As a local specialty industry, navel orange has been growing in scale with the continuous improvement of industrial technology. After years of development, Hubei Yichang, Jiangxi Gannan, Hunan Xinning and Chongqing Fengjie have vigorously planted navel oranges and established corresponding digital orchards, which has effectively promoted the development of local economy. Mastering the spatial distribution pattern and scale of navel oranges is of great significance to promoting local industrial economy, characteristic agricultural development and agricultural land development. However, navel oranges are very prone to diseases and pests. Under the stress of water and nutrients, the production of navel oranges is ultimately damaged, which greatly affects the economic income of local navel orange planting and hinders the market benefits of navel oranges. Based on this, it is necessary to master the occurrence period and characteristics of navel orange diseases and pests, monitor diseases and pests during the key growth period of navel oranges, formulate practical and feasible comprehensive prevention and control technologies, continuously improve the prevention and control effect, reduce the impact of adverse factors such as diseases and pests on the growth of navel oranges, and provide technical assistance and support for the increase of navel orange production and value.

[0003] A method and system for measuring crop growth parameters based on unmanned aerial vehicle multispectral remote sensing images, with publication number CN114202675A, comprises: dividing a crop area to be measured into a plurality of plots; collecting multispectral images using an unmanned aerial vehicle equipped with a multispectral remote sensing image acquisition device; correcting and processing the multispectral images to obtain a six-band multispectral orthophoto image; collecting measured data of growth parameters of crops in the area corresponding to the multispectral image; constructing an inversion model of growth parameters of crops in the area to be measured based on spectral reflectance and spectral index; and obtaining real-time non-destructive monitoring of growth parameters of crops in a large area through inversion of multispectral remote sensing images based on the model.

[0004] At present, most of the existing pest and disease monitoring methods are based on unmanned aerial vehicle remote sensing equipped with hyperspectral sensors to perform feature optimization and classification of diseased and healthy vegetation. This method has high time cost and cannot be promoted on a large scale. At the same time, the recognition efficiency and accuracy are low. Summary of the invention

[0005] In view of this, the present invention proposes a method for identifying navel orange pests and diseases, which can identify a large range of navel orange planting areas through multispectral image time series, and combine high-resolution satellites for on-board processing, which can improve the identification efficiency and accuracy of navel orange pests and diseases.

[0006] The technical solution of the present invention is implemented as follows: The present invention provides a method for identifying navel orange pests and diseases, comprising the following steps:

[0007] S1, obtaining multispectral satellite image information of the first satellite, the second satellite and the third satellite during the key growth period of navel orange, and preprocessing the multispectral satellite image information;

[0008] S2, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite and the third satellite to obtain a multispectral satellite synthetic image of each time window during the key growth period of navel orange;

[0009] S3, calculate the vegetation index for a multispectral satellite synthetic image in each time window to obtain the time series vegetation index;

[0010] S4, obtaining the field sampling sample information of navel orange and the time series vegetation index of the multi-spectral satellite synthetic image of the same period, drawing the vegetation index time series curve of the field sampling samples and the image interpretation samples of the same period, and screening to obtain the sample data set;

[0011] S5, inputting the screened sample data set into a random forest model for training to obtain a navel orange classification model, inputting a multispectral satellite synthetic image during the key growth period of navel oranges, using the navel orange classification model to classify the multispectral satellite synthetic image into navel oranges in the study area, and obtaining a spatial distribution map of navel oranges in the study area;

[0012] S6, based on the fourth satellite, obtain the satellite images in the study area, calculate the navel orange pest extraction index and gray level co-occurrence matrix on the satellite, extract the navel orange pest identification features, transmit the satellite images and pest identification feature set to the ground station, and splice them, use the navel orange spatial distribution map of the study area for mask processing, and obtain the navel orange filtered satellite image;

[0013] S7, a pest and disease identification model was constructed based on the support vector machine model. According to the pest and disease identification characteristics of navel orange and the navel orange filtered satellite image, the pest and disease identification model was used to identify diseased trees.

[0014] On the basis of the above technical solution, preferably, the step S1 of obtaining the multispectral satellite image information of the first satellite, the second satellite and the third satellite during the key growth period of navel orange and preprocessing the multispectral satellite image information includes the following sub-steps:

[0015] S11, based on the GEE cloud computing platform, preset a cloud ratio threshold and a cloud-snow coverage area threshold, and select satellite image information of the first satellite, the second satellite, and the third satellite whose cloud amount is less than the cloud ratio threshold, whose cloud-snow coverage area is less than the cloud-snow coverage area threshold, and whose location is within the key growth period of navel oranges;

[0016] S12, perform cloud removal and reflectivity standardization processing on each satellite image.

[0017] Based on the above technical solution, preferably, in step S12, each satellite image is subjected to cloud removal and reflectivity standardization processing respectively, wherein:

[0018] The reflectivity of the first satellite image is standardized and the expression is:

[0019] Blue s2 =0.0003+0.9570*S2Blue;

[0020] Green s2 =0.0015+1.0304*S2Green;

[0021] Red s2 =0.0041+0.9533*S2Red;

[0022] Nir s2 =0.0139+1.0157*S2Nir;

[0023] Sir1 s2 =0.0034+0.9522*S2Sir1;

[0024] Sir2 s2 =0.0004+0.9711*S2Sir2;

[0025] In the formula, S2Blue is the blue band of the first satellite data, S2Green is the green band of the first satellite data, S2Red is the red band of the first satellite data, S2Nir is the near infrared band of the first satellite data, S2Sir1 is the shortwave infrared band 1 of the first satellite data, S2Sir2 is the shortwave infrared band 2 of the first satellite data, Blue s2 The blue band is the first satellite reflectivity normalized, Green s2 The green band is the first satellite reflectivity normalized, Red s2 The red band is the normalized reflectivity of the first satellite, Nir s2 is the near-infrared band after the reflectivity of the first satellite is standardized, Sir1 s2 The shortwave infrared band 1, Sir2 after the normalization of the reflectivity of the first satellite s2 The shortwave infrared band 2 is the normalized reflectivity of the first satellite;

[0026] The reflectivity of the second satellite image is standardized and the expression is:

[0027] Blue L7 =0.0003+0.8474*L7Blue;

[0028] Green L7 =0.0088+0.8483*L7Green;

[0029] Red L7 =0.0061+0.9047*L7Red;

[0030] Nir L7 =0.0412+0.8462*L7Nir;

[0031] Sir1 L7 =0.0254+0.8937*L7Sir1;

[0032] Sir2 L7 =0.0172+0.9071*L7Sir2;

[0033] Wherein, L7Blue is the blue band of the second satellite data, L7Green is the green band of the second satellite data, L7Red is the red band of the second satellite data, L7Nir is the near infrared band of the second satellite data, L7Sir1 is the shortwave infrared band 1 of the second satellite data, L7Sir2 is the shortwave infrared band 2 of the second satellite data, Blue L7 The blue band is the normalized reflectivity of the second satellite, Green L7 The green band is the normalized reflectivity of the second satellite, Red L7 The red band is the normalized reflectivity of the second satellite, Nir L7 is the near-infrared band after the reflectivity of the second satellite is normalized, Sir1 L7 The shortwave infrared band 1, Sir2, after normalization of the reflectivity of the second satellite L7 The shortwave infrared band 2 is normalized for the reflectivity of the second satellite;

[0034] The reflectivity of the third satellite image is standardized and the expression is:

[0035] Blue L8 =0.0000275*L8Blue-0.2;

[0036] Green L8 =0.0000275*L8Green-0.2;

[0037] Red L8 =0.0000275*L8Red-0.2;

[0038] Nir L8 =0.0000275*L8Nir-0.2;

[0039] Sir1L8 =0.0000275*L8Sir1-0.2;

[0040] Sir2 L8 =0.0000275*L8Sir2-0.2;

[0041] Wherein, L8Blue is the blue band of the third satellite data, L8Green is the green band of the third satellite data, L8Red is the red band of the third satellite data, L8Nir is the near infrared band of the third satellite data, L8Sir1 is the shortwave infrared band 1 of the third satellite data, L8Sir2 is the shortwave infrared band 2 of the third satellite data, Blue L8 The blue band is the normalized reflectivity of the third satellite, Green L8 The green band is the standardized reflectivity of the third satellite, Red L8 The red band after the third satellite reflectivity standardization, Nir L8 is the near-infrared band after the reflectivity of the third satellite is standardized, Sir1 L8 The shortwave infrared band 1, Sir2 after the third satellite reflectivity is standardized L8 It is the shortwave infrared band 2 after the third satellite reflectivity is standardized.

[0042] On the basis of the above technical solution, preferably, the preset time window of fixed length in step S2 performs image fusion on the multispectral satellite image information of the first satellite, the second satellite and the third satellite based on the time window of fixed length, so as to obtain a multispectral satellite synthetic image of each time window in the key growth period of navel orange, which includes the following sub-steps:

[0043] S21, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite, and the third satellite corresponding to the time window, wherein:

[0044] The standardized blue band fusion calculation expression is:

[0045]

[0046] In the formula, Blue f is the blue band after standardized fusion of multi-spectral satellites, i is the number of satellite image data scenes of the first satellite in the corresponding time window, is the blue value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weight; n is the number of satellite image data scenes of the second satellite in the corresponding time window, Blue is the satellite image data scene of the nth second satellite in the corresponding time windowL7 The corresponding weight, z is the number of satellite image data scenes of the third satellite in the corresponding time window, is the blue field of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0047] The standardized green band fusion calculation expression is:

[0048]

[0049] In the formula, Green f It is the green band after standardized fusion of multi-spectral satellite. Green is the satellite image data scene of the first satellite of the i-th time window s2 The corresponding weights, Green is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Green is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0050] The standardized red band fusion calculation expression is:

[0051]

[0052] In the formula, Red f It is the red band after standardized fusion of multi-spectral satellite. is the Red value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weights, is the Red value of the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, is the Red value of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0053] The calculation expression of the standardized near-infrared band fusion is:

[0054]

[0055] In the formula, Nir f It is the red band after standardized fusion of multi-spectral satellite. Nir is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Nir is the satellite image data scene of the nth second satellite in the corresponding time windowL7 The corresponding weights, Nir is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0056] The standardized shortwave infrared band 1 fusion calculation expression is:

[0057]

[0058] In the formula, Sir1 f It is the red band after standardized fusion of multi-spectral satellite. Sir1 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Sir1 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Sir1 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0059] The standardized shortwave infrared band 2 fusion calculation expression is:

[0060]

[0061] In the formula, Sir2 f It is the red band after standardized fusion of multi-spectral satellite. Sir2 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Sir2 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Sir2 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0062] S22, crop the research area of ​​the fused multispectral satellite image to obtain a multispectral satellite composite image of each time window during the key growth period of navel orange.

[0063] On the basis of the above technical solution, preferably, the vegetation index calculation is performed on a multispectral satellite synthetic image of each time window in step S3 to obtain a time series vegetation index data set, wherein the vegetation index calculation includes the calculation of the normalized vegetation index NDVI, the ratio vegetation index RVI, the land surface water index LSWI, the enhanced vegetation index EVI and the greenness index VIGREEN, and the calculation expression is:

[0064] NDVI=(Nir f –Red f ) / (Nir f +Red f );

[0065] RVI=Nir f / Red f ;

[0066] LSWI=(Nir f -Sir1 f ) / (Nir f +Sir2 f );

[0067] EVI=(2.5*(Sir1 f -Red f )) / (Nir f +6*Red f –7.5*Blue f +1);

[0068] VIGREEN=(Green f –Red f ) / (Green f +Red f )).

[0069] On the basis of the above technical solution, preferably, the step S4 of obtaining the field sampling sample information of navel orange and the time series vegetation index of the multi-spectral satellite synthetic image of the same period, drawing the vegetation index time series curve of the field sampling sample and the image interpretation sample of the same period, and screening to obtain the sample data set includes the following sub-steps:

[0070] S31, obtaining field sampling sample points of different types of navel orange landforms, calculating the time series vegetation index according to the field sampling sample points, and drawing the normalized difference vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples;

[0071] S32, based on the GEE cloud computing platform, obtain the multispectral satellite synthetic images of the same period in the study area, and calculate the time series vegetation index. According to the normalized vegetation index NDVI and land surface water index LSWI vegetation index in the time series vegetation index dataset, draw the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the image interpretation samples of the same period, and compare them with the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples;

[0072] S33, using the time series curves of the normalized difference vegetation index NDVI and the land surface water index LSWI vegetation index of the field samples to screen the vegetation index time series curves of the interpreted samples, retaining the sample points with similar curves, and removing the sample points with inconsistent curves, and obtaining the sample data set after screening.

[0073] On the basis of the above technical scheme, preferably, the step S6 described in which, based on the fourth satellite, the satellite image in the study area is obtained, the navel orange pest extraction index and the gray level co-occurrence matrix are calculated on the satellite, and the pest identification features of the navel orange are extracted, the satellite image and the pest identification feature set are transmitted to the ground station, and spliced, and the spatial distribution map of the navel orange in the study area is used for mask processing to obtain the navel orange filtered satellite image, includes the following sub-steps:

[0074] S61, based on the fourth satellite, obtains on-board satellite images within the study area;

[0075] S62, calculating on-orbit a navel orange pest and disease extraction index and a gray-level symbiosis matrix on the satellite, wherein the navel orange pest and disease extraction index includes a greenness index VIGREEN and an atmospheric resistance index VARI;

[0076] S63, judging the greenness index VIGREEN, the atmospheric resistance index VARI and the gray-level symbiosis matrix pixel by pixel, presetting the greenness index threshold and the atmospheric resistance index threshold, if the greenness index VIGREEN is less than the greenness index threshold and the atmospheric resistance index VARI is less than the atmospheric resistance index threshold, assigning the pixel a label of 1, indicating that the pixel has potential pests and diseases; if the greenness index VIGREEN is greater than the greenness index threshold and the atmospheric resistance index VARI is greater than the atmospheric resistance index threshold, assigning the pixel a label of 0, indicating that the pixel has no pests and diseases; if the two judgment conditions are not met, assigning the pixel a label of 2, indicating that it is impossible to determine whether the pixel has pests and diseases, associating the pixel label value with the corresponding pixel point, and constructing a pixel label data set, using the greenness index VIGREEN, the atmospheric resistance index VARI, the gray-level symbiosis matrix and the pixel label data set as pest and disease identification features of navel orange;

[0077] S64, the satellite images and pest identification features of navel orange in the study area were transmitted to the ground station and mosaicked, and the spatial distribution map of navel orange in the study area was used for masking to remove interfering vegetation.

[0078] On the basis of the above technical solution, preferably, the pest and disease identification model is constructed based on the support vector machine model in step S7, and the pest and disease identification model is used to identify diseased trees according to the pest and disease identification features of navel orange and the navel orange filtered satellite image, wherein the following sub-steps are included:

[0079] S71, obtaining characteristic information of diseased and healthy navel orange tree samples on site, and inputting the characteristic information into a support vector machine model for iterative training to obtain a pest and disease recognition model;

[0080] S72, based on the navel orange pest and disease identification characteristics and the navel orange filtered satellite image, the pest and disease identification model is used to perform classification and identification, and a navel orange diseased tree identification result set is obtained;

[0081] S73, statistically analyzing the distribution of the greenness index VIGREEN and the atmospheric resistance index VARI of the navel orange disease and insect pest identification characteristics for the navel orange disease and insect pest identification result set, obtaining the greenness index VIGREEN and the atmospheric resistance index VARI values ​​of all healthy navel orange trees, and calculating the average values ​​of the pending greenness index VIGREEN and the pending atmospheric resistance index VARI respectively;

[0082] S74, updating the average value of the pending greenness index VIGREEN and the average value of the pending anti-atmospheric index VARI as the greenness index threshold and the anti-atmospheric index threshold on the satellite respectively.

[0083] In a second aspect, the present invention further provides an electronic device, comprising: at least one processor, at least one memory, a communication bus, a user interface, and a network interface;

[0084] Wherein, the processor, memory, user interface, and network interface communicate with each other via the bus;

[0085] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the above-mentioned navel orange disease and insect pest identification method.

[0086] In a third aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the above-mentioned navel orange pest and disease identification method.

[0087] The navel orange pest identification method of the present invention has the following beneficial effects compared with the prior art:

[0088] (1) By acquiring multispectral satellite image information from the first, second and third satellites during the critical growth period of navel oranges and performing image fusion on the multispectral satellite image information, a multispectral satellite composite image of each time window during the critical growth period of navel oranges is obtained, and the time series vegetation index is calculated. A navel orange classification model is constructed based on random forests, and the navel orange area is extracted from the multispectral satellite composite image to obtain the spatial distribution map of navel oranges in the study area, providing basic data support for the identification of navel orange pests and diseases. In addition, the high-resolution information from the fourth satellite is used to combine the high spatial resolution information and the rich spectral information of the multispectral satellite data to identify navel orange diseased trees, thereby obtaining the distribution information of diseased trees for agricultural workers to prevent and control navel orange pests and diseases. This method can realize the identification of a large range of navel orange planting areas, and combined with high-resolution satellites, it can improve the efficiency and accuracy of identifying navel orange pests and diseases.

[0089] (2) By fusing the multispectral satellite image information of the first satellite, the second satellite and the third satellite, the noise in the original image is removed, thereby obtaining a clearer and more informative image, which is conducive to improving the classification accuracy of navel oranges;

[0090] (3) By using the mean synthesis method within the time window to calculate the average value of each band, when the average reflectance of the pixel in each band is calculated, a weight value is added to adjust the impact of different data scenes on the final result, which can reduce the errors or anomalies that may be caused by a single observation value and improve the stability and reliability of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0092] Figure 1 It is a flow chart of the navel orange pest and disease identification method of the present invention;

[0093] Figure 2 It is a flowchart of the navel orange pest and disease identification method of the present invention;

[0094] Figure 3 The present invention is a schematic diagram of the navel orange growth period and the period of occurrence of diseases and insect pests in the navel orange disease and insect pest identification method. DETAILED DESCRIPTION

[0095] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0096] like Figure 1 As shown, the device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0097] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation of the device. In actual applications, the device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0098] like Figure 1 As shown, the memory 1005 as a medium may include an operating system, a network communication module, a user interface module, and a navel orange pest and disease identification method program.

[0099] exist Figure 1 In the device shown, the network interface 1004 is mainly used to establish a communication connection between the device and a server storing all data required in the navel orange pest and disease identification method system; the user interface 1003 is mainly used to interact with the user for data; the processor 1001 and the memory 1005 in the navel orange pest and disease identification method device of the present invention can be set in the navel orange pest and disease identification method device, and the navel orange pest and disease identification method device calls the navel orange pest and disease identification method program stored in the memory 1005 through the processor 1001, and executes the navel orange pest and disease identification method provided by the implementation of the present invention.

[0100] like Figure 2-3 As shown, a method for identifying diseases and insect pests of navel orange of the present invention comprises the following steps:

[0101] S1, obtaining multispectral satellite image information of the first satellite, the second satellite and the third satellite during the key growth period of navel orange, and preprocessing the multispectral satellite image information;

[0102] Step S1 includes the following sub-steps:

[0103] S11, based on the GEE cloud computing platform, preset a cloud ratio threshold and a cloud-snow coverage area threshold, and select satellite image information of the first satellite, the second satellite, and the third satellite whose cloud amount is less than the cloud ratio threshold, whose cloud-snow coverage area is less than the cloud-snow coverage area threshold, and whose location is within the key growth period of navel orange.

[0104] It should be noted that the first satellite is Sentinel-2, the second satellite is Landsat-7, and the third satellite is Landsat-8. The phenological information of navel orange is collected through literature to obtain the key growing period of navel orange. The key growing period of navel orange is from July to December each year. The multispectral satellite image data during the key growing period of navel orange is queried. Based on the GEE cloud computing platform, the satellite images of Sentinel-2, Landsat7 and Landsat8 with cloud cover less than 20% and cloud and snow coverage area less than 30% from July to December are selected.

[0105] S12, perform cloud removal and reflectivity standardization processing on each satellite image.

[0106] Among them, the reflectivity of the first satellite image is standardized, and the expression is:

[0107] Blue s2 =0.0003+0.9570*S2Blue;

[0108] Green s2 =0.0015+1.0304*S2Green;

[0109] Red s2 =0.0041+0.9533*S2Red;

[0110] Nir s2 =0.0139+1.0157*S2Nir;

[0111] Sir1 s2 =0.0034+0.9522*S2Sir1;

[0112] Sir2 s2=0.0004+0.9711*S2Sir2;

[0113] In the formula, S2Blue is the blue band of the first satellite data, S2Green is the green band of the first satellite data, S2Red is the red band of the first satellite data, S2Nir is the near infrared band of the first satellite data, S2Sir1 is the shortwave infrared band 1 of the first satellite data, S2Sir2 is the shortwave infrared band 2 of the first satellite data, Blue s2 The blue band is the first satellite reflectivity normalized, Green s2 The green band is the first satellite reflectivity normalized, Red s2 The red band is the normalized reflectivity of the first satellite, Nir s2 is the near-infrared band after the reflectivity of the first satellite is standardized, Sir1 s2 The shortwave infrared band 1, Sir2 after the normalization of the reflectivity of the first satellite s2 The shortwave infrared band 2 is the normalized reflectivity of the first satellite;

[0114] The reflectivity of the second satellite image is standardized and the expression is:

[0115] Blue L7 =0.0003+0.8474*L7Blue;

[0116] Green L7 =0.0088+0.8483*L7Green;

[0117] Red L7 =0.0061+0.9047*L7Red;

[0118] Nir L7 =0.0412+0.8462*L7Nir;

[0119] Sir1 L7 =0.0254+0.8937*L7Sir1;

[0120] Sir2 L7 =0.0172+0.9071*L7Sir2;

[0121] Wherein, L7Blue is the blue band of the second satellite data, L7Green is the green band of the second satellite data, L7Red is the red band of the second satellite data, L7Nir is the near infrared band of the second satellite data, L7Sir1 is the shortwave infrared band 1 of the second satellite data, L7Sir2 is the shortwave infrared band 2 of the second satellite data, Blue L7The blue band is the normalized reflectivity of the second satellite, Green L7 The green band is the normalized reflectivity of the second satellite, Red L7 The red band is the normalized reflectivity of the second satellite, Nir L7 is the near-infrared band after the reflectivity of the second satellite is normalized, Sir1 L7 The shortwave infrared band 1, Sir2, after normalization of the reflectivity of the second satellite L7 The shortwave infrared band 2 is normalized for the reflectivity of the second satellite;

[0122] The reflectivity of the third satellite image is standardized and the expression is:

[0123] Blue L8 =0.0000275*L8Blue-0.2;

[0124] Green L8 =0.0000275*L8Green-0.2;

[0125] Red L8 =0.0000275*L8Red-0.2;

[0126] Nir L8 =0.0000275*L8Nir-0.2;

[0127] Sir1 L8 =0.0000275*L8Sir1-0.2;

[0128] Sir2 L8 =0.0000275*L8Sir2-0.2;

[0129] Wherein, L8Blue is the blue band of the third satellite data, L8Green is the green band of the third satellite data, L8Red is the red band of the third satellite data, L8Nir is the near infrared band of the third satellite data, L8Sir1 is the shortwave infrared band 1 of the third satellite data, L8Sir2 is the shortwave infrared band 2 of the third satellite data, Blue L8 The blue band is the normalized reflectivity of the third satellite, Green L8 The green band is the standardized reflectivity of the third satellite, Red L8 The red band after the third satellite reflectivity standardization, Nir L8 is the near-infrared band after the reflectivity of the third satellite is standardized, Sir1 L8 The shortwave infrared band 1, Sir2 after the third satellite reflectivity is standardized L8 It is the shortwave infrared band 2 after the third satellite reflectivity is standardized.

[0130] S2, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite and the third satellite to obtain a multispectral satellite synthetic image of each time window during the key growth period of navel orange;

[0131] The following sub-steps are included:

[0132] S21, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite, and the third satellite corresponding to the time window, wherein:

[0133] The standardized blue band fusion calculation expression is:

[0134]

[0135] In the formula, Blue f is the blue band after standardized fusion of multi-spectral satellites, i is the number of satellite image data scenes of the first satellite in the corresponding time window, is the blue value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weight; n is the number of satellite image data scenes of the second satellite in the corresponding time window, Blue is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weight, z is the number of satellite image data scenes of the third satellite in the corresponding time window, is the blue field of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0136] The standardized green band fusion calculation expression is:

[0137]

[0138] In the formula, Green f It is the green band after standardized fusion of multi-spectral satellite. Green is the satellite image data scene of the first satellite of the i-th time window s2 The corresponding weights, Green is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Green is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0139] The standardized red band fusion calculation expression is:

[0140]

[0141] In the formula, Red f It is the red band after standardized fusion of multi-spectral satellite. is the Red value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weight, is the Red value of the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weight, is the Red value of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0142] The calculation expression of the standardized near-infrared band fusion is:

[0143]

[0144] In the formula, Nir f It is the red band after standardized fusion of multi-spectral satellite. Nir is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weight, Nir is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weight, Nir is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0145] The standardized shortwave infrared band 1 fusion calculation expression is:

[0146]

[0147] In the formula, Sir1 f It is the red band after standardized fusion of multi-spectral satellite. Sir1 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weight, Sir1 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weight, Sir1 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0148] The standardized shortwave infrared band 2 fusion calculation expression is:

[0149]

[0150] In the formula, Sir2 f It is the red band after standardized fusion of multi-spectral satellite. Sir2 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Sir2 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Sir2 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights;

[0151] S22, crop the research area of ​​the fused multispectral satellite image to obtain a multispectral satellite composite image of each time window during the key growth period of navel orange.

[0152] Specifically, the fixed step length of the time window in this embodiment is 20 days.

[0153] It should be noted that the average value of each band is calculated using the mean synthesis method within the time window. When the average reflectance of the pixel in each band is calculated, a weight value is added to adjust the impact of different data scenes on the final result. The weight value can be set by itself. The weight of the corresponding band reflectance can be determined based on factors such as data quality, reliability, and correlation in time series. By calculating the average reflectance of the pixel in each band, the error or anomaly that may be caused by a single observation value can be reduced, and the stability and reliability of the data can be improved. At the same time, by setting a time window of fixed length, the dynamic changes of the objects can be better monitored and analyzed.

[0154] S3, calculate the vegetation index for a multispectral satellite synthetic image in each time window to obtain the time series vegetation index;

[0155] The vegetation index calculation includes the calculation of the normalized difference vegetation index NDVI, the ratio vegetation index RVI, the land surface water index LSWI, the enhanced vegetation index EVI and the greenness index VIGREEN. The calculation expression is:

[0156] NDVI=(Nir f –Red f ) / (Nir f +Red f );

[0157] RVI=Nir f / Red f ;

[0158] LSWI=(Nir f -Sir1 f ) / (Nir f +Sir2 f );

[0159] EVI=(2.5*(Sir1 f -Red f )) / (Nir f +6*Red f –7.5*Blue f +1);

[0160] VIGREEN=(Green f –Red f ) / (Green f +Red f )).

[0161] Vegetation index can characterize the characteristics of vegetation. Its value is different in different growth stages of vegetation, so it can be used to distinguish the types of vegetation.

[0162] S4, obtaining the field sampling sample information of navel orange and the time series vegetation index of the multi-spectral satellite synthetic image of the same period, drawing the vegetation index time series curve of the field sampling samples and the image interpretation samples of the same period, and screening to obtain the sample data set;

[0163] The following sub-steps are included:

[0164] S31, obtaining field sampling sample points of different types of navel orange landforms, calculating the time series vegetation index according to the field sampling sample points, and drawing the normalized difference vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples;

[0165] S32, based on the GEE cloud computing platform, obtain the multispectral satellite synthetic images of the same period in the study area, and calculate the time series vegetation index. According to the normalized vegetation index NDVI and land surface water index LSWI vegetation index in the time series vegetation index dataset, draw the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the image interpretation samples of the same period, and compare them with the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples;

[0166] S33, using the time series curves of the normalized difference vegetation index NDVI and the land surface water index LSWI vegetation index of the field samples to screen the vegetation index time series curves of the interpreted samples, retaining the sample points with similar curves, and removing the sample points with inconsistent curves, and obtaining the sample data set after screening.

[0167] It should be noted that field sampling sample points of navel orange, water bodies, urban construction land, cultivated land, forest land and other land types were obtained, and the NDVI and LSWI vegetation index time series curves of the field sampling samples were drawn on Google Earth software. Manual interpretation was performed by referring to the image comparison of Google Earth at the same time, and sample points of navel orange, water bodies, urban construction land, cultivated land, forest land and other land types were added. The NDVI and LSWI vegetation index time series curves of the manually interpreted sample points were drawn and compared with the field sampling samples for sample screening. The manually interpreted samples were screened with the NDVI and LSWI vegetation index time series curves of the field samples, and the sample points with similar curves were retained by manual selection, and the sample points with completely inconsistent curves were kicked out.

[0168] S5, input the screened sample data set into the random forest model for training to obtain a navel orange classification model, input the multispectral satellite synthetic image during the key growth period of the navel orange, use the navel orange classification model to classify the multispectral satellite synthetic image into the navel oranges in the study area, and obtain the spatial distribution map of the navel oranges in the study area.

[0169] It should be noted that, of the sample data set after screening, 80% of the sample points are used for random forest model training, and 20% of the sample points are used for validation set for accuracy verification. In order to balance the calculation time and prediction accuracy, the total number of trees in the random forest, ntree, is set to 500, the mtry parameter is the square root of the total number of input features, the loss function selects the out-of-bag error rate, ooberror, and other parameters remain at the default values. When the training accuracy of the random forest model reaches 90%, the model training is considered successful. Using this model, the validation sample set is input, and the confusion matrix, mapping accuracy PA, user accuracy UA and overall accuracy OA are calculated to evaluate the navel orange recognition effect. The higher the value, the better the accuracy.

[0170] in,

[0171]

[0172]

[0173]

[0174] Where N is the total number of samples used for accuracy evaluation; X ii is the number of correctly classified pixels in the i-th category of pixels in the classification result; X ij is the total number of pixels of the i-th category in the classification result; X ji is the total number of pixels of the i-th category in the actual situation.

[0175] In this embodiment, 55 navel orange sample points, 2 water bodies, 12 urban building land, 33 cultivated land, 14 woodland and 25 other land use types of field sampling sample points are obtained, and the NDVI and LSWI vegetation index time series curves of the field sampling samples are drawn. On the Google Earth software, the Google Earth image comparison of the same period is used for manual interpretation, and 40 navel oranges, 10 water bodies, 10 urban building land, 30 cultivated land, 30 woodland and 30 other land use types of sample points are added. The NDVI and LSWI vegetation index time series curves of the manually interpreted sample points are drawn and compared with the field sampling samples for sample screening. The manually interpreted samples are screened with the time series curves of the NDVI and LSWI vegetation index of the field samples, and the human eye selects the sample points with similar curves and keeps them, and kicks out the sample points with completely inconsistent curves. Before screening, there were a total of 95 sample points of navel orange, 12 water bodies, 22 urban building land, 63 cultivated land, 44 forest land, and 55 other land use types. After screening, there were a total of 88 samples of navel orange, 12 water bodies, 22 urban building land, 58 cultivated land, and 44 other land use types. After screening, 80% of the sample points were used for random forest model training and 20% of the sample points were used as validation sets for accuracy verification. In order to balance the calculation time and prediction accuracy, the total number of trees in the random forest, ntree, was set to 500, the mtry parameter was the square root of the total number of input features, which was 3, and other parameters remained at the default values. The training accuracy of the random forest model reached 92.3%, and the model training was considered successful. Using this model, input the validation sample set, calculate the confusion matrix, and use the validation sample set to calculate the confusion matrix and mapping accuracy PA, user accuracy UA and overall accuracy OA to evaluate the navel orange recognition effect. The UA accuracy of navel orange is 77.27%, the PA accuracy is 78.30%, the UA accuracy of water body is 98.75%, the PA accuracy is 96.82%, the UA accuracy of urban building land is 95.84%, the PA accuracy is 89.34%, the UA accuracy of forest land is 97.32%, the PA accuracy is 96.77%, the UA accuracy of cultivated land is 81.48%, the PA accuracy is 91.28%, and the UA accuracy of other land types is 80.97%, the PA accuracy is 75.97%. The overall accuracy is 91.15%. The time series vegetation index dataset and land feature type data of the sample points are used as the input data of the random forest model to train the navel orange random forest model. Finally, the navel orange random forest model is used to classify the study area and obtain the spatial distribution map of navel orange.

[0176] S6, based on the fourth satellite, obtain the satellite images in the study area, calculate the navel orange pest extraction index and gray level co-occurrence matrix on the satellite, extract the navel orange pest identification features, transmit the satellite images and pest identification feature set to the ground station, and splice them, use the navel orange spatial distribution map of the study area for mask processing, and obtain the navel orange filtered satellite image;

[0177] The following sub-steps are included:

[0178] S61, based on the fourth satellite, obtains on-board satellite images within the study area;

[0179] S62, calculating on-orbit a navel orange pest and disease extraction index and a gray-level symbiosis matrix on the satellite, wherein the navel orange pest and disease extraction index includes a greenness index VIGREEN and an atmospheric resistance index VARI;

[0180] S63, judging the greenness index VIGREEN, the atmospheric resistance index VARI and the gray-level symbiosis matrix pixel by pixel, presetting the greenness index threshold and the atmospheric resistance index threshold, if the greenness index VIGREEN is less than the greenness index threshold and the atmospheric resistance index VARI is less than the atmospheric resistance index threshold, assigning the pixel a label of 1, indicating that the pixel has potential pests and diseases; if the greenness index VIGREEN is greater than the greenness index threshold and the atmospheric resistance index VARI is greater than the atmospheric resistance index threshold, assigning the pixel a label of 0, indicating that the pixel has no pests and diseases; if the two judgment conditions are not met, assigning the pixel a label of 2, indicating that it is impossible to determine whether the pixel has pests and diseases, associating the pixel label value with the corresponding pixel point, and constructing a pixel label data set, using the greenness index VIGREEN, the atmospheric resistance index VARI, the gray-level symbiosis matrix and the pixel label data set as pest and disease identification features of navel orange;

[0181] S64, the satellite images and pest identification features of navel orange in the study area were transmitted to the ground station and mosaicked, and the spatial distribution map of navel orange in the study area was used for masking to remove interfering vegetation.

[0182] Specifically, the fourth satellite in this embodiment is the Luojia-3 01 satellite, wherein the Luojia-3 satellite is an intelligent surveying and remote sensing experimental satellite that can acquire Bayer format images with a spatial resolution better than 0.75 meters, and has a high-performance real-time processing hardware platform and a software platform that supports dynamic loading and opening. It can realize on-orbit real-time intelligent processing of remote sensing data, greatly improving the efficiency of high-resolution remote sensing data processing.

[0183] S7, a pest and disease identification model was constructed based on the support vector machine model. According to the pest and disease identification characteristics of navel orange and the navel orange filtered satellite image, the pest and disease identification model was used to identify diseased trees.

[0184] The following sub-steps are included:

[0185] S71, obtaining characteristic information of diseased and healthy navel orange tree samples on site, and inputting the characteristic information into a support vector machine model for iterative training to obtain a pest and disease recognition model;

[0186] S72, based on the navel orange pest and disease identification characteristics and the navel orange filtered satellite image, the pest and disease identification model is used to perform classification and identification, and a navel orange diseased tree identification result set is obtained;

[0187] S73, statistically analyzing the distribution of the greenness index VIGREEN and the atmospheric resistance index VARI of the navel orange disease and insect pest identification characteristics for the navel orange disease and insect pest identification result set, obtaining the greenness index VIGREEN and the atmospheric resistance index VARI values ​​of all healthy navel orange trees, and calculating the average values ​​of the pending greenness index VIGREEN and the pending atmospheric resistance index VARI respectively;

[0188] S74, updating the average value of the pending greenness index VIGREEN and the average value of the pending anti-atmospheric index VARI as the greenness index threshold and the anti-atmospheric index threshold on the satellite respectively.

[0189] Among them, the support vector machine model uses Gaussian kernel function and optimization iteration to adjust the penalty coefficient, and other parameters remain default to obtain the navel orange diseased tree identification results. The overall verification accuracy of the classification is 81%. The diseased trees are classified and identified through the disease and pest identification model, so as to obtain the diseased tree distribution information for agricultural workers to carry out navel orange disease and pest control. The undetermined green index VIGREEN average value and the undetermined atmospheric resistance index VARI average value are used as the green index threshold and the atmospheric resistance index threshold and are updated synchronously on the satellite, so as to achieve real-time navel orange disease and pest identification both on and off the satellite. In the absence of navel orange disease and pest samples, on-orbit identification can be achieved based on the green index threshold and the atmospheric resistance index threshold. The data processing and calculation of a Luojia-3 image is limited to within 5 minutes, and the identification process is transformed from the original cumbersome data downloading, reflectivity calibration, disease and pest feature calculation and identification to an efficient and real-time on-orbit monitoring technology, which greatly improves the efficiency of monitoring and identification.

[0190] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, wherein the computer instructions enable the computer to implement all or part of the steps of the method described in the embodiment of the present invention. The storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk, and other media that can store program codes.

[0191] The system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be distributed to multiple network units. A person skilled in the art may select some or all of the modules according to actual needs to achieve the purpose of the solution of this embodiment without creative effort.

[0192] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying navel orange pests and diseases, characterized in that: The following steps are involved: S1, obtaining multispectral satellite image information of the first satellite, the second satellite and the third satellite during the key growth period of navel orange, and preprocessing the multispectral satellite image information; S2, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite and the third satellite to obtain a multispectral satellite synthetic image of each time window during the key growth period of navel orange; S3, calculate the vegetation index for a multispectral satellite synthetic image in each time window to obtain the time series vegetation index; S4, obtaining the field sampling sample information of navel orange and the time series vegetation index of the multi-spectral satellite synthetic image of the same period, drawing the vegetation index time series curve of the field sampling samples and the image interpretation samples of the same period, and screening to obtain the sample data set; S5, inputting the screened sample data set into a random forest model for training to obtain a navel orange classification model, inputting a multispectral satellite synthetic image during the key growth period of navel oranges, using the navel orange classification model to classify the multispectral satellite synthetic image into navel oranges in the study area, and obtaining a spatial distribution map of navel oranges in the study area; S6, based on the fourth satellite, obtain the satellite images in the study area, calculate the navel orange pest extraction index and gray level co-occurrence matrix on the satellite, extract the navel orange pest identification features, transmit the satellite images and pest identification feature set to the ground station, and splice them, use the navel orange spatial distribution map of the study area for mask processing, and obtain the navel orange filtered satellite image; S7, a pest and disease identification model was constructed based on the support vector machine model. According to the pest and disease identification characteristics of navel orange and the navel orange filtered satellite images, the pest and disease identification model was used to identify diseased trees.

2. The method for identifying navel orange pests and diseases as claimed in claim 1, characterized in that: Step S1 is to obtain multispectral satellite image information of the first satellite, the second satellite and the third satellite during the key growth period of navel oranges, and pre-process the multispectral satellite image information, which includes the following sub-steps: S11, based on the GEE cloud computing platform, preset a cloud ratio threshold and a cloud-snow coverage area threshold, and select satellite image information of the first satellite, the second satellite, and the third satellite whose cloud amount is less than the cloud ratio threshold, whose cloud-snow coverage area is less than the cloud-snow coverage area threshold, and whose location is within the key growth period of navel oranges; S12, perform cloud removal and reflectivity standardization processing on each satellite image.

3. The navel orange pest identification method according to claim 2, characterized in that: In step S12, each satellite image is subjected to cloud removal and reflectivity standardization processing, wherein: The reflectivity of the first satellite image is standardized and the expression is: Blue s2 =0.0003+0.9570*S2Blue; Green s2 =0.0015+1.0304*S2Green; Red s2 =0.0041+0.9533*S2Red; Nir s2 =0.0139+1.0157*S2Nir; Sir1 s2 =0.0034+0.9522*S2Sir1; Sir2 s2 =0.0004+0.9711*S2Sir2; In the formula, S2Blue is the blue band of the first satellite data, S2Green is the green band of the first satellite data, S2Red is the red band of the first satellite data, S2Nir is the near infrared band of the first satellite data, S2Sir1 is the shortwave infrared band 1 of the first satellite data, S2Sir2 is the shortwave infrared band 2 of the first satellite data, Blue s2 The blue band is the first satellite reflectivity normalized, Green s2 The green band is the first satellite reflectivity normalized, Red s2 The red band is the normalized reflectivity of the first satellite, Nir s2 is the near-infrared band after the reflectivity of the first satellite is standardized, Sir1 s2 The shortwave infrared band 1, Sir2 after the normalization of the reflectivity of the first satellite s2 The shortwave infrared band 2 is the normalized reflectivity of the first satellite; The reflectivity of the second satellite image is standardized and the expression is: Blue L7 =0.0003+0.8474*L7Blue; Green L7 =0.0088+0.8483*L7Green; Red L7 =0.0061+0.9047*L7Red; Nir L7 =0.0412+0.8462*L7Nir; Cheese1 L7 =0.0254+0.8937*L7Sir1; Sir2 L7 =0.0172+0.9071*L7Sir2; Wherein, L7Blue is the blue band of the second satellite data, L7Green is the green band of the second satellite data, L7Red is the red band of the second satellite data, L7Nir is the near infrared band of the second satellite data, L7Sir1 is the shortwave infrared band 1 of the second satellite data, L7Sir2 is the shortwave infrared band 2 of the second satellite data, Blue L7 The blue band is the normalized reflectivity of the second satellite, Green L7 The green band is the normalized reflectivity of the second satellite, Red L7 The red band is the normalized reflectivity of the second satellite, Nir L7 is the near-infrared band after the reflectivity of the second satellite is normalized, Sir1 L7 The shortwave infrared band 1, Sir2, after normalization of the reflectivity of the second satellite L7 The shortwave infrared band 2 is normalized for the reflectivity of the second satellite; The reflectivity of the third satellite image is standardized and the expression is: Blue L8 =0.0000275*L8Blue-0.2; Green L8 =0.0000275*L8Green-0.2; Red L8 =0.0000275*L8Red-0.2; Nir L8 =0.0000275*L8Nir-0.2; Sir1 L8 =0.0000275*L8Sir1-0.2; Sir2 L8 =0.0000275*L8Sir2-0.2; Wherein, L8Blue is the blue band of the third satellite data, L8Green is the green band of the third satellite data, L8Red is the red band of the third satellite data, L8Nir is the near infrared band of the third satellite data, L8Sir1 is the shortwave infrared band 1 of the third satellite data, L8Sir2 is the shortwave infrared band 2 of the third satellite data, Blue L8 The blue band is the normalized reflectivity of the third satellite, Green L8 The green band is the standardized reflectivity of the third satellite, Red L8 The red band after the third satellite reflectivity standardization, Nir L8 is the near-infrared band after the reflectivity of the third satellite is standardized, Sir1 L8 The shortwave infrared band 1, Sir2 after the third satellite reflectivity is standardized L8 It is the shortwave infrared band 2 after the third satellite reflectivity is standardized.

4. The method for identifying navel orange pests and diseases as claimed in claim 3, characterized in that: The time window of fixed length is preset in step S2, and based on the time window of fixed length, the multispectral satellite image information of the first satellite, the second satellite and the third satellite is image fused to obtain a multispectral satellite synthetic image of each time window in the key growth period of navel orange, which includes the following sub-steps: S21, presetting a time window of fixed length, and based on the time window of fixed length, performing image fusion on the multispectral satellite image information of the first satellite, the second satellite, and the third satellite corresponding to the time window, wherein: The standardized blue band fusion calculation expression is: In the formula, Blue f is the blue band after standardized fusion of multi-spectral satellites, i is the number of satellite image data scenes of the first satellite in the corresponding time window, is the blue value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weight; n is the number of satellite image data scenes of the second satellite in the corresponding time window, Blue is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weight, z is the number of satellite image data scenes of the third satellite in the corresponding time window, is the blue field of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; The standardized green band fusion calculation expression is: In the formula, Green f It is the green band after standardized fusion of multi-spectral satellite. Green is the satellite image data scene of the first satellite of the i-th time window s2 The corresponding weights, Green is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Green is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; The standardized red band fusion calculation expression is: In the formula, Red f It is the red band after standardized fusion of multi-spectral satellite. is the Red value of the satellite image data scene of the first satellite of the i-th satellite in the corresponding time window s2 The corresponding weights, is the Red value of the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, is the Red value of the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; The calculation expression of the standardized near-infrared band fusion is: In the formula, Nir f It is the red band after standardized fusion of multi-spectral satellite. Nir is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Nir is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Nir is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; The standardized shortwave infrared band 1 fusion calculation expression is: In the formula, Sir1 f It is the red band after standardized fusion of multi-spectral satellite. Sir1 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Sir1 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Sir1 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; The standardized shortwave infrared band 2 fusion calculation expression is: In the formula, Sir2 f It is the red band after standardized fusion of multi-spectral satellite. Sir2 is the satellite image data scene of the i-th first satellite in the corresponding time window S2 The corresponding weights, Sir2 is the satellite image data scene of the nth second satellite in the corresponding time window L7 The corresponding weights, Sir2 is the satellite image data scene of the zth third satellite in the corresponding time window L8 The corresponding weights; S22, crop the research area of ​​the fused multispectral satellite image to obtain a multispectral satellite composite image of each time window during the key growth period of navel orange.

5. The navel orange pest identification method according to claim 4, characterized in that: In step S3, the vegetation index is calculated for a multispectral satellite synthetic image in each time window to obtain a time series vegetation index data set, wherein the vegetation index calculation includes the calculation of the normalized vegetation index NDVI, the ratio vegetation index RVI, the land surface water index LSWI, the enhanced vegetation index EVI and the greenness index VIGREEN, and the calculation expression is: NDVI=(Nir f –Red f ) / (Nir f +Red f ); RVI=Nir f / Red f , LSWI=(Nir f -Sir1 f ) / (Nir f +Sir2 f ); EVI=(2.5*(Sir1 f -Red f )) / (Nir f +6*Red f –7.5*Blue f +1); VIGREEN=(Green f –Red f ) / (Green f +Red f ))。 6. The method for identifying navel orange pests and diseases as claimed in claim 5, characterized in that: Step S4 is to obtain the field sampling sample information of navel orange and the time series vegetation index of the multi-spectral satellite synthetic image of the same period, draw the vegetation index time series curve of the field sampling sample and the image interpretation sample of the same period, and screen to obtain the sample data set, which includes the following sub-steps: S31, obtaining field sampling sample points of different types of navel orange landforms, calculating the time series vegetation index according to the field sampling sample points, and drawing the normalized difference vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples; S32, based on the GEE cloud computing platform, obtain the multispectral satellite synthetic images of the same period in the study area, and calculate the time series vegetation index. According to the normalized vegetation index NDVI and land surface water index LSWI vegetation index in the time series vegetation index dataset, draw the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the image interpretation samples of the same period, and compare them with the normalized vegetation index NDVI and land surface water index LSWI vegetation index time series curves of the field sampling samples; S33, using the time series curves of the normalized difference vegetation index NDVI and the land surface water index LSWI vegetation index of the field samples to screen the vegetation index time series curves of the interpreted samples, retaining the sample points with similar curves, and removing the sample points with inconsistent curves, and obtaining the sample data set after screening.

7. The method for identifying navel orange pests and diseases as claimed in claim 6, characterized in that: The method described in step S6 is to obtain satellite images on the satellite within the study area based on the fourth satellite, calculate the navel orange pest and disease extraction index and gray level co-occurrence matrix on the satellite, extract the navel orange pest and disease identification features, transmit the satellite images on the satellite and the pest and disease identification feature set to the ground station, and perform splicing, and perform mask processing using the navel orange spatial distribution map of the study area to obtain the navel orange filtered satellite image, including the following sub-steps: S61, based on the fourth satellite, obtains on-board satellite images within the study area; S62, calculating on-orbit a navel orange pest and disease extraction index and a gray-level symbiosis matrix on the satellite, wherein the navel orange pest and disease extraction index includes a greenness index VIGREEN and an atmospheric resistance index VARI; S63, judging the greenness index VIGREEN, the atmospheric resistance index VARI and the gray-level symbiosis matrix pixel by pixel, presetting the greenness index threshold and the atmospheric resistance index threshold, if the greenness index VIGREEN is less than the greenness index threshold and the atmospheric resistance index VARI is less than the atmospheric resistance index threshold, assigning the pixel a label of 1, indicating that the pixel has potential pests and diseases; if the greenness index VIGREEN is greater than the greenness index threshold and the atmospheric resistance index VARI is greater than the atmospheric resistance index threshold, assigning the pixel a label of 0, indicating that the pixel has no pests and diseases; if the two judgment conditions are not met, assigning the pixel a label of 2, indicating that it is impossible to determine whether the pixel has pests and diseases, associating the pixel label value with the corresponding pixel point, and constructing a pixel label data set, using the greenness index VIGREEN, the atmospheric resistance index VARI, the gray-level symbiosis matrix and the pixel label data set as pest and disease identification features of navel orange; S64, the satellite images and pest identification features of navel orange in the study area were transmitted to the ground station and mosaicked, and the spatial distribution map of navel orange in the study area was used for masking to remove interfering vegetation.

8. The method for identifying navel orange pests and diseases as claimed in claim 7, characterized in that: The step S7 constructs a pest and disease identification model based on the support vector machine model, and uses the pest and disease identification model to identify diseased trees according to the pest and disease identification features of navel oranges and the navel orange filtered satellite image, which includes the following sub-steps: S71, obtaining characteristic information of diseased and healthy navel orange tree samples on site, and inputting the characteristic information into a support vector machine model for iterative training to obtain a pest and disease recognition model; S72, based on the navel orange pest and disease identification characteristics and the navel orange filtered satellite image, the pest and disease identification model is used to perform classification and identification, and a navel orange diseased tree identification result set is obtained; S73, statistically analyzing the distribution of the greenness index VIGREEN and the atmospheric resistance index VARI of the navel orange disease and insect pest identification characteristics for the navel orange disease and insect pest identification result set, obtaining the greenness index VIGREEN and the atmospheric resistance index VARI values ​​of all healthy navel orange trees, and calculating the average values ​​of the pending greenness index VIGREEN and the pending atmospheric resistance index VARI respectively; S74, updating the average value of the pending greenness index VIGREEN and the average value of the pending anti-atmospheric index VARI as the greenness index threshold and the anti-atmospheric index threshold on the satellite respectively.

9. An electronic device, characterized in that: include: at least one processor, at least one memory, a communication bus, a user interface, and a network interface; Wherein, the processor, memory, user interface, and network interface communicate with each other via the bus; The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the navel orange pest and disease identification method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the navel orange pest and disease identification method according to any one of claims 1 to 8.

Images