Wine grape gray mold grade discrimination method and device based on unmanned aerial vehicle remote sensing, and medium

Abstract

The invention discloses a wine grape gray mold grade discrimination method and device based on unmanned aerial vehicle remote sensing and a medium, and relates to the technical field of unmanned aerial vehicle remote sensing. The method comprises the following steps: preprocessing unmanned aerial vehicle image data to obtain a multi-period hyperspectral image; extracting region-of-interest spectral data from the multi-period hyperspectral image based on the geographic position data; carrying out vegetation index feature extraction and physiological and biochemical feature extraction on the spectroscopic data of the region of interest; the vegetation index features and the physiological and biochemical features are combined with the original spectral band to be connected according to a time sequence to form a multi-temporal feature vector; correlation analysis is carried out on the multi-temporal feature vectors, and target feature vectors most relevant to gray mold are screened out; according to the method, the grape gray mold grade is taken as a target feature, a disease monitoring model based on a time sequence is constructed, training prediction is carried out by utilizing a target feature vector, and a grape gray mold monitoring model is obtained, so that efficient, accurate and high-adaptability wine grape gray mold grade judgment is realized.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle remote sensing technology, and more specifically, to a method, device and medium for distinguishing the gray mold grade of wine grapes based on unmanned aerial vehicle remote sensing. Background Art

[0002] Wine grapes are one of the most important economic crops in the world. However, as a common fungal disease, gray mold seriously affects the yield and quality of grapes. Traditional gray mold monitoring methods mainly rely on manual field surveys, which are not only time-consuming and labor-intensive, but also limited by human subjective judgment and prone to errors. With the development of remote sensing technology, drone remote sensing has gradually become an emerging tool for agricultural disease monitoring, especially the introduction of hyperspectral imaging technology, which provides new possibilities for accurate disease monitoring.

[0003] At present, drone-based disease monitoring technology has gradually been applied to the agricultural field, especially in disease monitoring. Research has made some progress. Most of the studies have used drones equipped with visible light and multi-spectral cameras to take aerial photos of vineyards to obtain high-resolution spectral data. By building a spectral feature library and combining machine learning algorithms (such as support vector machines, random forests, etc.), gray mold can be classified and graded.

[0004] However, existing methods usually rely on data collection at a single time point, which makes it difficult to reflect the dynamic development of the disease. The recognition accuracy is low in the early stages of the disease or in areas where symptoms are not obvious, and the model generalization ability is limited. Therefore, there is an urgent need for an efficient, accurate, and adaptable disease grade discrimination method to achieve real-time monitoring and precise management of grape gray mold. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method, device and medium for distinguishing the gray mold grade of wine grapes based on unmanned aerial vehicle remote sensing, so as to achieve efficient, accurate and adaptable gray mold grade distinction of wine grapes.

[0006] In a first aspect, the present invention provides a method for distinguishing the grade of gray mold of wine grapes based on unmanned aerial vehicle remote sensing, the method comprising:

[0007] Preprocess the drone image data to obtain multi-period hyperspectral images;

[0008] Acquire manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographic location data;

[0009] Extracting spectral data of a region of interest from the multi-period hyperspectral images based on the geographic location data;

[0010] Extracting vegetation index features from the spectral data of the region of interest, and constructing vegetation index features that are sensitive to grape gray mold;

[0011] Extracting physiological and biochemical characteristics from the spectral data of the region of interest;

[0012] Connecting the vegetation index feature and the physiological and biochemical feature in combination with the original spectral bands in time sequence to form a multi-temporal feature vector;

[0013] Performing correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to Botrytis cinerea;

[0014] Taking the grape gray mold disease level in the corresponding image area collected on the ground as the target feature, a disease monitoring model based on time series is constructed, and the target feature vector is input into the disease monitoring model based on time series for training and prediction, so as to obtain a grape gray mold disease monitoring model based on drone images.

[0015] Furthermore, the drone image data is preprocessed to obtain multi-period hyperspectral images, including:

[0016] Obtaining drone hyperspectral images of wine grape disease areas over multiple periods

[0017] Perform lens calibration, import raw data, dark current data and lens correction files, and export lens calibrated data;

[0018] Perform reflectivity calibration, import the lens calibration data, dark current data and white frame data obtained from the standard white board, and export the reflectivity calibration data;

[0019] Perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data;

[0020] Perform image stitching on the final data to output an orthophoto;

[0021] The orthophotos are imported into ArcGIS, and the multi-period images are geo-referenced using ground image control points to obtain multi-period hyperspectral images.

[0022] Furthermore, the vegetation index characteristics include normalized difference vegetation index, photochemical reflectance index, red edge stress vegetation index and chlorophyll absorption ratio index.

[0023] Furthermore, the vegetation index feature is extracted from the spectral data of the region of interest by the following formulas (1) to (4), and the vegetation index feature sensitive to grape gray mold is constructed:

[0024]

[0025]

[0026] In the formula, NDVI represents the normalized difference vegetation index, PRI represents the photochemical reflectance index, RVSI represents the red edge vegetation stress index, CARI represents the chlorophyll absorption ratio index, and r 800 represents the spectral reflectance at 800nm, r 670 represents the spectral reflectance at 670nm, r 531 represents the spectral reflectance at 531nm, r 570 represents the spectral reflectance at 570nm, r 714 represents the spectral reflectance at 714nm, r 752 represents the spectral reflectance at 752nm, r 733 represents the spectral reflectance at 733nm, r 700 represents the spectral reflectance at 700nm, r 550 Indicates the spectral reflectance at 550nm.

[0027] Furthermore, the physiological and biochemical characteristics of the spectral data of the region of interest are extracted, including:

[0028] Generate simulated spectral lookup tables for wine grapes using the PROSAIL model;

[0029] The corresponding physiological and biochemical characteristics were inferred using the lookup table and spectral data.

[0030] Furthermore, the physiological and biochemical characteristics include chlorophyll, carotenoids, nitrogen, dry matter, leaf water content, anthocyanin, leaf area index and leaf inclination angle.

[0031] In a second aspect, the present invention provides a device for distinguishing the gray mold grade of wine grapes based on drone remote sensing, the device comprising:

[0032] A data preprocessing unit is configured to preprocess the drone image data to obtain multi-period hyperspectral images;

[0033] A data acquisition unit is configured to acquire manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographical location data;

[0034] A first extraction unit is configured to extract spectral data of a region of interest from the multi-period hyperspectral image based on the geographic location data;

[0035] A second extraction unit is configured to extract vegetation index features from the spectral data of the region of interest, and construct vegetation index features that are sensitive to grape gray mold;

[0036] A third extraction unit is configured to extract physiological and biochemical features from the spectral data of the region of interest;

[0037] A time series connection unit is configured to connect the vegetation index feature and the physiological and biochemical feature in combination with the original spectral band in time series to form a multi-temporal feature vector;

[0038] A feature screening unit is configured to perform correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to gray mold;

[0039] The model building unit is configured to build a disease monitoring model based on time series with the disease level of grape gray mold in the corresponding image area collected on the ground as the target feature, input the target feature vector into the disease monitoring model based on time series for training and prediction, and obtain a grape gray mold disease monitoring model based on drone images.

[0040] Furthermore, the data preprocessing unit is further configured to:

[0041] Obtaining drone hyperspectral images of wine grape disease areas over multiple periods

[0042] Perform lens calibration, import raw data, dark current data and lens correction files, and export lens calibrated data;

[0043] Perform reflectivity calibration, import the lens calibration data, dark current data and white frame data obtained from the standard white board, and export the reflectivity calibration data;

[0044] Perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data;

[0045] Perform image stitching on the final data to output an orthophoto;

[0046] The orthophotos are imported into ArcGIS, and the multi-period images are geo-referenced using ground image control points to obtain multi-period hyperspectral images.

[0047] Furthermore, the vegetation index characteristics include normalized difference vegetation index, photochemical reflectance index, red edge stress vegetation index and chlorophyll absorption ratio index.

[0048] In a third aspect, the present invention provides a readable storage medium, wherein the readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method as described above.

[0049] The present invention has at least the following beneficial effects:

[0050] The present invention uses multi-temporal UAV hyperspectral images, takes into account the temporal characteristics of the disease spectrum, and achieves accurate estimation over a large range. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The whole flow chart of a method for distinguishing the gray mold grade of wine grapes based on UAV remote sensing according to an embodiment of the present invention is shown.

[0052] Figure 2 A flowchart of preprocessing drone image data according to an embodiment of the present invention is shown;

[0053] Figure 3 A schematic diagram of splicing multi-phase feature vectors according to an embodiment of the present invention is shown;

[0054] Figure 4 A structural diagram of a device for distinguishing the gray mold grade of wine grapes based on drone remote sensing according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0055] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and specific embodiments, but are not intended to limit the present invention. For the various steps described herein, if there is no necessity for a causal relationship between each other, the order in which they are described as examples herein should not be regarded as a limitation, and those skilled in the art should know that they can be adjusted in order, as long as the logic between them is not destroyed, resulting in the inability to implement the entire process.

[0056] The embodiment of the present invention provides a method for distinguishing the grade of gray mold of wine grapes based on unmanned aerial vehicle remote sensing. Figure 1 As shown, the method for distinguishing the gray mold grade of wine grapes based on UAV remote sensing includes the following steps S10 to S80.

[0057] S10, preprocessing the UAV image data to obtain multi-period hyperspectral images.

[0058] In some embodiments, Figure 2 As shown, step S10 is implemented by the following steps S11 to S16.

[0059] S11, obtain drone hyperspectral images of wine grape disease areas in multiple periods.

[0060] S12, perform lens calibration, import UAV hyperspectral images of wine grape disease areas in multiple periods, dark current data, and lens calibration files, and export the lens calibrated data.

[0061] S13, perform reflectivity calibration, import the lens calibrated data, dark current data, and white frame data obtained from a standard white board, and export the reflectivity calibrated data.

[0062] S14, perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data.

[0063] S15, import the final data into Photoscan, perform image stitching, and output orthophotos.

[0064] S16, import the orthophotos into ArcGIS, use ground image control points to georeference the multi-period images, and obtain multi-period hyperspectral images.

[0065] S20, obtaining manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographical location data.

[0066] S30, extracting spectral data of a region of interest from the multi-period hyperspectral images based on the geographic location data.

[0067] In this embodiment, ArcGIS is used to extract spectral data of the region of interest from the multi-period hyperspectral image calibrated in step S10 according to the geographical location data in step S20.

[0068] S40, extracting vegetation index features from the spectral data of the region of interest, and constructing vegetation index features that are sensitive to grape botrytis cinerea.

[0069] In some embodiments, the vegetation index features include normalized difference vegetation index (NDVI), photochemical reflectance index (PRI), red edge stress vegetation index (RVSI), and chlorophyll absorption ratio index (CARI). The calculation formulas are as follows: formula (1) to formula (4).

[0070]

[0071] In the formula, NDVI represents the normalized difference vegetation index, PRI represents the photochemical reflectance index, RVSI represents the red edge vegetation stress index, CARI represents the chlorophyll absorption ratio index, and r 800 represents the spectral reflectance at 800nm, r 670represents the spectral reflectance at 670nm, r 531 represents the spectral reflectance at 531nm, r 570 represents the spectral reflectance at 570nm, r 714 represents the spectral reflectance at 714nm, r 752 represents the spectral reflectance at 752nm, r 733 represents the spectral reflectance at 733nm, r 700 represents the spectral reflectance at 700nm, r 550 Indicates the spectral reflectance at 550nm.

[0072] S50, extracting physiological and biochemical characteristics from the spectral data of the region of interest.

[0073] In some embodiments, the multi-period spectral data in step S30 is subjected to physiological and biochemical feature extraction. First, the PROSAIL model is used to generate a simulated spectral lookup table suitable for wine grapes, and then the lookup table and spectral data are used to infer the corresponding physiological and biochemical characteristics, including chlorophyll (Cab), carotenoids (Car), nitrogen (N), dry matter (Cm), leaf water content (Cw), anthocyanin (Anth), leaf area index (LAI), and leaf inclination angle (LAD).

[0074] S60, combining the vegetation index feature and the physiological and biochemical feature with the original spectral bands in time sequence to form a multi-temporal feature vector.

[0075] In this embodiment, Figure 3 As shown, the vegetation index features extracted in step S40 and the physiological and biochemical features extracted in step S50 are combined with the original spectral bands and connected in time sequence to form a multi-temporal feature vector.

[0076] S70, performing correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to Botrytis cinerea.

[0077] S80, taking the grape gray mold disease level of the corresponding image area collected on the ground as the target feature, constructing a disease monitoring model based on time series, inputting the target feature vector into the disease monitoring model based on time series for training and prediction, and obtaining a grape gray mold disease monitoring model based on drone images.

[0078] In this embodiment, the disease monitoring model based on time series can be an LSTM network, and the grape gray mold disease monitoring model based on drone images can be finally obtained to realize the gray mold grade discrimination of wine grapes based on drone remote sensing images.

[0079] The embodiment of the present invention also provides a device for distinguishing the gray mold grade of wine grapes based on drone remote sensing, such as Figure 4 As shown, the device comprises:

[0080] The data preprocessing unit 401 is configured to preprocess the drone image data to obtain multi-period hyperspectral images;

[0081] The data acquisition unit 402 is configured to acquire manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographical location data;

[0082] A first extraction unit 403 is configured to extract spectral data of a region of interest from the multi-period hyperspectral image based on the geographic location data;

[0083] The second extraction unit 404 is configured to extract vegetation index features from the spectral data of the region of interest, and construct vegetation index features that are sensitive to grape gray mold;

[0084] The third extraction unit 405 is configured to extract physiological and biochemical features from the spectral data of the region of interest;

[0085] A time series connection unit 406 is configured to connect the vegetation index feature and the physiological and biochemical feature in combination with the original spectral band in time series to form a multi-temporal feature vector;

[0086] A feature screening unit 407 is configured to perform correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to gray mold;

[0087] The model building unit 408 is configured to build a disease monitoring model based on time series with the disease level of grape gray mold in the corresponding image area collected on the ground as the target feature, input the target feature vector into the disease monitoring model based on time series for training and prediction, and obtain a grape gray mold disease monitoring model based on drone images.

[0088] In some embodiments, the data pre-processing unit is further configured to:

[0089] Obtaining drone hyperspectral images of wine grape disease areas over multiple periods

[0090] Perform lens calibration, import raw data, dark current data and lens correction files, and export lens calibrated data;

[0091] Perform reflectivity calibration, import the lens calibration data, dark current data and white frame data obtained from the standard white board, and export the reflectivity calibration data;

[0092] Perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data;

[0093] Perform image stitching on the final data to output an orthophoto;

[0094] The orthophotos are imported into ArcGIS, and the multi-period images are geo-referenced using ground image control points to obtain multi-period hyperspectral images.

[0095] In some embodiments, the vegetation index features include normalized difference vegetation index, photochemical reflectance index, red edge stress vegetation index and chlorophyll absorption ratio index.

[0096] In some embodiments, the second extraction unit is further configured to extract vegetation index features from the spectral data of the region of interest by using the following formulas (1) to (4) to construct vegetation index features that are sensitive to grape gray mold:

[0097]

[0098] In the formula, NDVI represents the normalized difference vegetation index, PRI represents the photochemical reflectance index, RVSI represents the red edge vegetation stress index, CARI represents the chlorophyll absorption ratio index, and r 800 represents the spectral reflectance at 800nm, r 670 represents the spectral reflectance at 670nm, r 531 represents the spectral reflectance at 531nm, r 570 represents the spectral reflectance at 570nm, r 714 represents the spectral reflectance at 714nm, r 752 represents the spectral reflectance at 752nm, r 733 represents the spectral reflectance at 733nm, r 700 represents the spectral reflectance at 700nm, r 550 Indicates the spectral reflectance at 550nm.

[0099] In some embodiments, the third extraction unit is further configured to:

[0100] Generate simulated spectral lookup tables for wine grapes using the PROSAIL model;

[0101] The corresponding physiological and biochemical characteristics were inferred using the lookup table and spectral data.

[0102] In some embodiments, the physiological and biochemical characteristics include chlorophyll, carotenoids, nitrogen, dry matter, leaf water content, anthocyanins, leaf area index, and leaf inclination.

[0103] It should be noted that the structures of the various wine grape gray mold grade discrimination devices based on drone remote sensing described in this embodiment belong to the same technical concept as the wine grape gray mold grade discrimination method based on drone remote sensing described previously, and achieve the same beneficial effects through the same principles, which will not be repeated here.

[0104] An embodiment of the present invention further provides a readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the method described in any of the above embodiments.

[0105] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. For example, those of ordinary skill in the art can use other embodiments when reading the above description. In addition, in the above-mentioned specific embodiments, various features can be grouped together to simplify the present invention. This should not be interpreted as a feature of an invention that is not claimed for protection being necessary for any claim. On the contrary, the subject matter of the present invention may be less than all the features of the embodiments of a specific invention. Thus, the following claims are incorporated into the specific embodiments as examples or embodiments, wherein each claim is independently used as a separate embodiment, and it is considered that these embodiments can be combined with each other in various combinations or arrangements. The scope of the present invention should be determined with reference to the full scope of the equivalent forms of the attached claims and these claims.

Claims

1. A method for distinguishing the grade of gray mold of wine grapes based on unmanned aerial vehicle remote sensing, characterized in that: The method comprises: Preprocess the drone image data to obtain multi-period hyperspectral images; Acquire manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographic location data; Extracting spectral data of a region of interest from the multi-period hyperspectral images based on the geographic location data; Extracting vegetation index features from the spectral data of the region of interest, and constructing vegetation index features that are sensitive to grape gray mold; Extracting physiological and biochemical characteristics from the spectral data of the region of interest; Connecting the vegetation index feature and the physiological and biochemical feature in combination with the original spectral bands in time sequence to form a multi-temporal feature vector; Performing correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to Botrytis cinerea; Taking the grape gray mold disease level in the corresponding image area collected on the ground as the target feature, a disease monitoring model based on time series is constructed, and the target feature vector is input into the disease monitoring model based on time series for training and prediction, so as to obtain a grape gray mold disease monitoring model based on drone images.

2. The method for distinguishing the grade of gray mold of wine grapes based on UAV remote sensing according to claim 1 is characterized in that: Preprocess the drone image data to obtain multi-period hyperspectral images, including: Acquire drone hyperspectral images of wine grape disease areas over multiple periods; Perform lens calibration, import raw data, dark current data and lens correction files, and export lens calibrated data; Perform reflectivity calibration, import the lens calibration data, dark current data and white frame data obtained from the standard white board, and export the reflectivity calibration data; Perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data; Perform image stitching on the final data to output an orthophoto; The orthophotos are imported into ArcGIS, and the multi-period images are geo-referenced using ground image control points to obtain multi-period hyperspectral images.

3. The method for distinguishing the grade of gray mold of wine grapes based on UAV remote sensing according to claim 1 is characterized in that: The vegetation index characteristics include normalized difference vegetation index, photochemical reflectance index, red edge stress vegetation index and chlorophyll absorption ratio index.

4. The method for distinguishing the gray mold grade of wine grapes based on UAV remote sensing according to claim 3 is characterized in that: The vegetation index feature is extracted from the spectral data of the region of interest by the following formulas (1) to (4), and the vegetation index feature sensitive to grape gray mold is constructed: In the formula, NDVI represents the normalized difference vegetation index, PRI represents the photochemical reflectance index, RVSI represents the red edge vegetation stress index, CARI represents the chlorophyll absorption ratio index, and r 800 represents the spectral reflectance at 800nm, r 670 represents the spectral reflectance at 670nm, r 531 represents the spectral reflectance at 531nm, r 570 represents the spectral reflectance at 570nm, r 714 represents the spectral reflectance at 714nm, r 752 represents the spectral reflectance at 752nm, r 733 represents the spectral reflectance at 733nm, r 700 represents the spectral reflectance at 700nm, r 550 Indicates the spectral reflectance at 550nm.

5. The method for distinguishing the grade of gray mold of wine grapes based on UAV remote sensing according to claim 1 is characterized in that: Extracting physiological and biochemical characteristics of the spectral data of the region of interest includes: Generate simulated spectral lookup tables for wine grapes using the PROSAIL model; The corresponding physiological and biochemical characteristics were inferred using the lookup table and spectral data.

6. The method for distinguishing the gray mold grade of wine grapes based on UAV remote sensing according to claim 5, characterized in that: The physiological and biochemical characteristics include chlorophyll, carotenoids, nitrogen, dry matter, leaf water content, anthocyanin, leaf area index and leaf inclination angle.

7. A device for distinguishing the gray mold grade of wine grapes based on drone remote sensing, characterized in that: The device comprises: A data preprocessing unit is configured to preprocess the drone image data to obtain multi-period hyperspectral images; A data acquisition unit is configured to acquire manual identification data of gray mold of wine grapes on the ground in the image area corresponding to the multi-period hyperspectral image and corresponding geographical location data; A first extraction unit is configured to extract spectral data of a region of interest from the multi-period hyperspectral image based on the geographic location data; A second extraction unit is configured to extract vegetation index features from the spectral data of the region of interest, and construct vegetation index features that are sensitive to grape gray mold; A third extraction unit is configured to extract physiological and biochemical features from the spectral data of the region of interest; A time series connection unit is configured to connect the vegetation index feature and the physiological and biochemical feature in combination with the original spectral band in time series to form a multi-temporal feature vector; A feature screening unit is configured to perform correlation analysis on the multi-phase feature vectors to screen out the target feature vector most relevant to gray mold; The model building unit is configured to build a disease monitoring model based on time series with the disease level of grape gray mold in the corresponding image area collected on the ground as the target feature, input the target feature vector into the disease monitoring model based on time series for training and prediction, and obtain a grape gray mold disease monitoring model based on drone images.

8. The device for distinguishing the gray mold grade of wine grapes based on UAV remote sensing according to claim 7 is characterized in that: The data preprocessing unit is further configured to: Obtaining drone hyperspectral images of wine grape disease areas over multiple periods Perform lens calibration, import raw data, dark current data and lens correction files, and export lens calibrated data; Perform reflectivity calibration, import the lens calibration data, dark current data and white frame data obtained from the standard white board, and export the reflectivity calibration data; Perform atmospheric correction, load the data after reflectivity calibration, select 50% standard reflectivity gray cloth, import the gray cloth calibration curve, perform batch processing, and output the final data; Perform image stitching on the final data to output an orthophoto; The orthophotos are imported into ArcGIS, and the multi-period images are geo-referenced using ground image control points to obtain multi-period hyperspectral images.

9. The device for distinguishing the gray mold grade of wine grapes based on UAV remote sensing according to claim 7, characterized in that: The vegetation index characteristics include normalized difference vegetation index, photochemical reflectance index, red edge stress vegetation index and chlorophyll absorption ratio index. 10 . A non-transitory computer-readable storage medium storing instructions, which, when executed by a processor, executes the method according to claim 1 .

Images