Peanut sample automatic generation method and system based on multi-source TWDTW

Abstract

The invention belongs to the technical field of agricultural remote sensing, and relates to a multi-source TWDTW-based peanut sample automatic generation method and system. The method comprises the following steps: establishing a standard phenological curve corresponding to each feature of crops in each phenological area; splicing to obtain a multi-source characteristic standard phenological curve; acquiring a multi-source characteristic phenological curve of the to-be-detected sample point; calculating an accumulated distance between the multi-source characteristic phenological curve and the multi-source characteristic standard phenological curve of each type of crops; determining the crop type of the to-be-detected sample point, and obtaining the sample data of the crop type. The invention provides an automatic peanut sample generation method combining multi-source remote sensing data and growth cycle characteristics, a multi-source characteristic standard phenological curve is established, an accumulated distance is calculated, and a machine learning method is utilized to improve the accuracy of peanut early recognition and reduce the difficulty and cost of peanut sample acquisition.

Description

Technical Field

[0001] The present invention belongs to the technical field of agricultural remote sensing, and more specifically, relates to a method and system for automatically generating peanut samples based on multi-source TWDTW. Background Art

[0002] Peanuts are important oil crops and cash crops. Timely and accurate peanut spatial distribution information is crucial for agricultural management, food security, and the prediction and decision-making of the oil crop market. Remote sensing is an effective tool for crop type mapping, and supervised classification methods such as machine learning and deep learning are the main methods for crop type mapping. However, the performance of these supervised classification methods highly depends on the quantity and quality of training samples. Although traditional field surveys provide the most direct and reliable means of sample collection, they are time-consuming and costly, especially for large-scale mapping of crops in the early stage. How to obtain sufficient samples in the early stage of the crop growing season is the key issue for improving the mapping accuracy and timeliness of peanuts.

[0003] Currently, the sample automatic generation technology combining remote sensing time series data and crop phenological periods has become an effective means to solve the problem of sample acquisition. However, most methods mainly utilize optical data features and conduct research on the extraction of major food crops (wheat and corn) separately. In complex planting areas and growing seasons, the phenomenon of different electromagnetic radiation spectral characteristics of crops is often more significant, which poses a challenge to the early identification of peanuts. In particular, the similarity in the growth periods between peanuts and root and tuber crops such as sweet potatoes increases the difficulty of identification. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method, system, and electronic device for automatically generating peanut samples based on multi-source TWDTW.

[0005] In a first aspect, the present invention provides a method for automatically generating peanut samples based on multi-source TWDTW, including: Establishing several phenological zones according to the area sizes of various types of crops in the region, and establishing standard phenological curves corresponding to each feature of the crops in each phenological zone according to the field survey sample point data and the preprocessed remote sensing image time series data; Stitching the standard phenological curves of different features in the same area and at the same time period into a multi-source feature standard phenological curve; Obtaining the land cover data of the point to be detected, randomly generating sample points within the cultivated land plot, calculating the time series data of the sample points to be detected, and obtaining the multi-source feature phenological curve of the sample points to be detected; Calculating the cumulative distance between the multi-source feature phenological curves of each type of crop and the multi-source feature standard phenological curve, and determining the threshold interval corresponding to the cumulative distance between the multi-source feature phenological curves of each type of crop and the multi-source feature standard phenological curve; Calculate the cumulative distance between the multi-source feature phenological curve of the sample point to be detected and the multi-source feature standard phenological curve; Determine the crop type of the sample point to be detected according to the magnitude of the cumulative distance and the threshold interval, and obtain the sample data of this crop type.

[0006] In a second aspect, the present invention provides a peanut sample automatic generation system based on multi-source TWDTW, including a standard phenological curve establishment unit, a splicing unit, a multi-source feature phenological curve generation unit, a first data processing unit, a second data processing unit and an output unit; The standard phenological curve establishment unit is used to establish several phenological regions according to the area sizes of various types of crops in the region, and establish the standard phenological curves corresponding to each feature of the crops in each phenological region according to the sample point data of on-site investigation and the preprocessed remote sensing image time series data; The splicing unit is used to splice the standard phenological curves of different features in the same region and at the same time period into a multi-source feature standard phenological curve; The multi-source feature phenological curve generation unit is used to obtain the land cover data of the point to be detected, randomly generate sample points within the cultivated land plot, calculate the time series data of the sample points to be detected, and obtain the multi-source feature phenological curve of the sample points to be detected; The first data processing unit is used to calculate the cumulative distance between the multi-source feature phenological curve of each type of crop and the multi-source feature standard phenological curve, and determine the threshold interval corresponding to the cumulative distance between the multi-source feature phenological curve and the multi-source feature standard phenological curve of each type of crop; The second data processing unit is used to calculate the cumulative distance between the multi-source feature phenological curve of the sample point to be detected and the multi-source feature standard phenological curve; The output unit is used to determine the crop type of the sample point to be detected according to the magnitude of the cumulative distance and the threshold interval, and obtain the sample data of this crop type.

[0007] On the basis of the above technical solutions, the present invention can also be improved as follows.

[0008] Further, the remote sensing image time series data includes SAR (Synthetic Aperture Radar) data, multi-spectral remote sensing data and land cover data.

[0009] Further, the preprocessing of the remote sensing image time series data of the crop includes: filtering the remote sensing image time series data, converting the coordinate system to the World Geodetic System, and resampling the azimuth and range resolutions of the image to 10m; using the S-G smoothing filtering algorithm to remove the time series smoothing thermal noise; radiometric correction; using DEM data for terrain correction.

[0010] Furthermore, the characteristics of the crops include band characteristics, texture characteristics, and vegetation index characteristics.

[0011] Furthermore, calculate the cumulative distance between the multi-source characteristic phenological curves and the multi-source characteristic standard phenological curves of each type of crop, including: Let the time series of the multi-source characteristic phenological curve of the crop be , with a length of , and the th characteristic be ; Let the time series of the multi-source characteristic standard phenological curve be , with a length of , and the th characteristic be ; Then: ; ; Let be the base distance between the th characteristic and the th characteristic, and be the distance matrix between the th characteristic and the th characteristic. Then: ; Let the time weight factor be , be the gain factor, be the distance factor. The distance factor is the difference in the time series lengths between the th characteristic and the th characteristic, , be the set ratio of the sequence time span. Then: ; ; Let be the node corresponding to the th characteristic and the th characteristic, be the node corresponding to the th characteristic and the th characteristic, be the node corresponding to the th characteristic and the th characteristic, be the node corresponding to the th characteristic and the th characteristic, and at the node has a cumulative distance of , find the shortest cumulative distance, and the previous node is , and the minimum value between ; Let the minimum cumulative distance path be L, then: .

[0012] Furthermore, when calculating the cumulative distance between the multi-source characteristic phenological curves of various types of crops and the multi-source characteristic standard phenological curves, let the cumulative distance of the th crop be , be the weight of the cumulative distance of the th crop, and the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region is , the weight of the cumulative distance of the th crop is determined by the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region, then: ; Let be the cumulative distance sum between the standard phenological curves in the th phenological region, be the weight of the cumulative distance of the th crop, then: .

[0013] Furthermore, according to the magnitude of the cumulative distance and the threshold interval, determine the crop type of the sample point to be detected, obtain the preliminary classification sample, and resample the time series of the point to be detected to be consistent with the crop standard time series.

[0014] Furthermore, calculate the Manhattan distance between each preliminary classification sample and the corresponding labeled crop. If the Manhattan distance is greater than the set threshold, retain the preliminary classification sample; otherwise, discard the preliminary classification sample; let the Manhattan distance be , be the eigenvalue of the time series of the point to be detected at time , be the eigenvalue of the crop standard time series at time , be the number of data in the time series, then: .

[0015] Furthermore, resample the time series of the point to be detected to be consistent with the crop standard time series before calculating the Manhattan distance between each preliminary classification sample and the corresponding labeled crop.

[0016] The beneficial effects of the present invention are as follows: The present invention proposes an automatic peanut sample generation method combining multi-source remote sensing data and growth cycle characteristics, establishes a multi-source feature standard phenological curve, calculates the cumulative distance between the multi-source feature phenological curve of the sample point to be detected and the multi-source feature standard phenological curve, and uses machine learning methods to improve the accuracy of early peanut recognition and reduce the difficulty and cost of obtaining peanut samples. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 FIG. is a schematic diagram of the automatic peanut sample generation method based on multi-source TWDTW provided in Embodiment 1 of the present invention; Figure 2 FIG. is a comparison diagram before and after S-G smoothing filtering; Figure 3 FIG. is a multi-source TWDTW partition diagram, where (a) in the appendix Figure 3 is the multi-source TWDTW partition diagram, and (b) in the appendix Figure 3 is the normalized vegetation index curve diagram of each crop, and (c) in the appendix Figure 3 is the contrast curve, and (d) in the appendix Figure 3 is the phenological curve of three stages of the crop growth period; Figure 4 FIG. is a schematic diagram of the automatic peanut sample generation system based on multi-source TWDTW provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations.

[0019] Embodiment 1 As an embodiment, as shown in the appendix Figure 1 To solve the above technical problems, the present embodiment provides an automatic peanut sample generation method based on multi-source TWDTW, including: Establish several phenological regions according to the area sizes of various types of crops in the region, and establish standard phenological curves corresponding to each feature of the crops in each phenological region according to the on-site survey sample point data and the preprocessed remote sensing image time series data; Stitch the standard phenological curves of different features in the same region and at the same time period into a multi-source feature standard phenological curve; Obtain the land cover data of the point to be detected, randomly generate sample points within the cultivated land plot, calculate the time series data of the sample points to be detected, and obtain the multi-source feature phenological curve of the sample points to be detected; Calculate the cumulative distance between the multi-source characteristic phenological curves of each type of crop and the multi-source characteristic standard phenological curve, and determine the threshold interval corresponding to the cumulative distance between the multi-source characteristic phenological curve and the multi-source characteristic standard phenological curve of each type of crop; Calculate the cumulative distance between the multi-source characteristic phenological curve of the sample point to be detected and the multi-source characteristic standard phenological curve; Determine the crop type of the sample point to be detected according to the magnitude of the cumulative distance and the threshold interval, and obtain the sample data of this crop type.

[0020] In the actual application process, the Sentinel-1 satellite ground range multi-look image set within the selected SAR data area is used. This image set includes three spatial resolutions of 10m, 25m, and 40m, four polarization band combinations, and three imaging modes. Only the images with the scanning mode of IW (Interferometric Wide Swath) and the polarization modes of VH (Vertical-Horizontal polarization) and VV (Vertical-Vertical polarization) need to be selected. The multi-spectral remote sensing data is Sentinel-2 MSI L2A data (Sentinel-2 Multi-Spectral Instrument Level-2A data). The bands required for the experiment are 10 bands in total, including visible light (B2~B4), red edge (B5~B7), near-infrared (B8 and B8A), and short-wave infrared (B11 and B12). The land cover data uses a 10m resolution global land cover map. The ground crop sample data is field sampling data of crops, including longitude and latitude information and crop type labels.

[0021] Optionally, the remote sensing image time series data includes SAR (Synthetic Aperture Radar) data, multi-spectral remote sensing data, and land cover data.

[0022] Optionally, the preprocessing of the remote sensing image time series data of crops includes: filtering the remote sensing image time series data, converting the coordinate system to the World Geodetic System, and resampling the azimuth and range resolutions of the image to 10m; using the S-G smoothing filter algorithm to remove the time series smoothing thermal noise; performing radiometric calibration; and performing terrain correction using DEM data.

[0023] In the actual application process, after obtaining the Sentinel-1 satellite remote sensing images in GEE (Google Earth Engine), in order to reduce or eliminate the error influence of speckle noise and retain the edge details of the images, the Lee-sigma algorithm is used to filter the Sentinel-1 satellite remote sensing images. At the same time, to facilitate the registration of the Sentinel-1 satellite remote sensing images and the heterologous images of multi-spectral remote sensing data, it is necessary to convert the coordinate system of the Sentinel-1 satellite remote sensing images into the World Geodetic System and resample the azimuth and range resolutions of the images to 10m. Due to the influence of the solar angle, observation angle, sensor sensitivity, bidirectional reflection of the ground objects, and aerosols, the Sentinel-1 satellite remote sensing images and multi-spectral remote sensing data often contain a lot of noise. Therefore, through the S-G smoothing filter (Savitzky-Golay Smoothing Filter) for temporal smoothing, it is easier to reflect the change trend reflected by the data itself, such as Figure 2 As shown, the horizontal axis is time, unit: day, the vertical axis is the normalized vegetation index, D is the original remote sensing image temporal data, L1 is the linear interpolation curve, and L2 is the data after S-G smoothing filtering.

[0024] Optionally, the characteristics of the crops include band characteristics, texture characteristics, and vegetation index characteristics.

[0025] The basic principle of the TWDTW algorithm (Time Weighted Danymic Time Warping) is to find the best alignment between time series, calculate the cumulative distance, and thus evaluate the similarity.

[0026] Optionally, calculating the cumulative distance between the multi-source characteristic phenological curves of various types of crops and the multi-source characteristic standard phenological curves includes: setting the time series of the multi-source characteristic phenological curve of the crop as , with a length of , and the th characteristic is ; setting the time series of the multi-source characteristic standard phenological curve as , with a length of , and the th characteristic is ; then: ; ; Let be the base distance between the th characteristic and the th characteristic, be the The distance matrix between the th feature, then: ; Let the time weight factor be , be the gain factor, be the distance factor, and the distance factor is the difference in the time series lengths between the th feature and the th feature, , is the set ratio of the sequence time span, then: ; ; Let be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, and at the node is the cumulative distance , find the shortest cumulative distance, and the previous node is , and is the minimum value between them, then: ; Let the path of the minimum cumulative distance be L, then: .

[0027] Optionally, when calculating the cumulative distance between the multi-source feature phenological curves of various types of crops and the multi-source feature standard phenological curves, let the cumulative distance of the th crop be , be the weight of the cumulative distance of the th crop, and the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region is , and the weight of the cumulative distance of the th crop is determined by the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region, then: ; Let be the cumulative distance sum between the standard phenological curves in the th phenological region, be the weight of the cumulative distance of the th crop, then: .

[0028] Let be the cumulative distance between corn and peanut, be the cumulative distance between corn and sweet potato, be the cumulative distance between peanut and sweet potato, be the total cumulative distance of the region, then: .

[0029] When calculating the cumulative distance between the standard phenological curves of different crops using the TWDTW algorithm, there are differences in the time when the cumulative distance of the standard phenological curves of different features is the largest.

[0030] As shown in the appendix Figure 3 , L1 represents peanut, L2 represents sweet potato, L3 represents corn. In the appendix Figure 3 , (a) is the multi-source TWDTW partition diagram. The horizontal axis is the crop growth time, unit: day, starting from March 1, 2024. The vertical axis is VV+VH (the sum of vertical-horizontal polarization and vertical-vertical polarization). In the appendix Figure 3 , (b) is the normalized vegetation index curve diagram of each crop. The horizontal axis is time, unit: day, and the vertical axis is the normalized vegetation index; in the appendix Figure 3 , (c) is the contrast curve. The horizontal axis is time, unit: day, and the vertical axis is the contrast calculated using the near-infrared band; in the appendix Figure 3 , (d) divides the crop growth period into three stages, and different feature phenological curves are used in each stage. Among them, the differences in the standard phenological curves of SAR image features are mainly concentrated in the pre-sowing period of crops (0-80 days), the differences in the normalized vegetation index features of optical images are concentrated in the crop growth and development period (80-130 days), and the differences in the texture features calculated using the near-infrared band of multi-spectral remote sensing data are mainly concentrated in the late growth period of crops (after 130 days).

[0031] To make full use of the existing remote sensing image data to improve the accuracy of sample generation, the present invention proposes to splice the standard phenological curves of different features into a multi-source feature standard phenological curve with greater differentiation based on the multi-source TWDTW algorithm according to the difference in feature sensitivity of crops in different growth periods, as shown in Figure 3 , (d). When generating samples, the to-be-matched phenological curves are also generated according to the forms of different features in different periods.

[0032] Optionally, determine the crop type of the sample points to be detected according to the cumulative distance and the threshold interval to obtain the initially classified samples, and resample the time series of the points to be detected to be consistent with the crop standard time series.

[0033] Optionally, calculate the Manhattan distance between each initially classified sample and the corresponding labeled crop. If the Manhattan distance is greater than the set threshold, retain the initially classified sample; otherwise, discard the initially classified sample. Let the Manhattan distance be , be the eigenvalue of the time series of the point to be detected at time , be the eigenvalue of the crop standard time series at time , be the number of data in the time series, then: .

[0034] Optionally, resample the time series of the points to be detected to be consistent with the crop standard time series before calculating the Manhattan distance between each initially classified sample and the corresponding labeled crop.

[0035] To verify the reliability of the present invention, use the field investigation points and the method of the present invention to generate samples for the target area, and perform an overlay analysis on the sample generation results and the peanut plot map of the first year. The generated sample points basically fall on the peanut plots, indicating that the method of the present invention has a high accuracy in generating samples. Further, use the existing field sampling point data of the second year to conduct a sample generation experiment. If the field sampling data is missing in this area, use the field sampling points of this area in previous years to replace it. The sample point results of generating sample data using limited field sampling points also prove that the method of the present invention has a high accuracy in generating samples.

[0036] The present invention proposes a method for automatically generating peanut samples by combining multi-source remote sensing data and growth cycle characteristics. By using SAR data, multi-spectral remote sensing data and land cover data, a multi-source feature standard phenological curve is established, the cumulative distance between the multi-source feature phenological curve and the multi-source feature standard phenological curve of the sample points to be detected is calculated, and a machine learning method is used to improve the accuracy of early peanut recognition and reduce the difficulty and cost of obtaining peanut samples.

[0037] Embodiment 2 Based on the same principle as the method shown in Embodiment 1 of the present invention, as shown in the appendix Figure 4 shown, the embodiment of the present invention also provides a system for automatically generating peanut samples based on multi-source TWDTW, including a standard phenological curve establishment unit, a splicing unit, a multi-source feature phenological curve generation unit, a first data processing unit, a second data processing unit and an output unit; A standard phenological curve establishment unit is used to establish several phenological zones according to the area sizes of various types of crops in a region, and establish standard phenological curves corresponding to each feature of the crops in each phenological zone based on the field survey sample point data and the preprocessed remote sensing image time series data; A splicing unit is used to splice the standard phenological curves of different features in the same area and the same time period into a multi-source feature standard phenological curve; A multi-source feature phenological curve generation unit is used to obtain the land cover data of the point to be detected, randomly generate sample points within the cultivated land plot, calculate the time series data of the sample points to be detected, and obtain the multi-source feature phenological curve of the sample points to be detected; A first data processing unit is used to calculate the cumulative distance between the multi-source feature phenological curves of various types of crops and the multi-source feature standard phenological curve, and determine the threshold interval corresponding to the cumulative distance between the multi-source feature phenological curves of various types of crops and the multi-source feature standard phenological curve; A second data processing unit is used to calculate the cumulative distance between the multi-source feature phenological curve of the sample points to be detected and the multi-source feature standard phenological curve; An output unit is used to determine the crop type of the sample points to be detected according to the magnitude of the cumulative distance and the threshold interval, and obtain the sample data of the crop type.

[0038] Optionally, the remote sensing image time series data includes SAR (Synthetic Aperture Radar) data, multi-spectral remote sensing data, and land cover data.

[0039] Optionally, the preprocessing of the remote sensing image time series data of the crops includes: filtering the remote sensing image time series data, converting the coordinate system to the World Geodetic System, and resampling the image azimuth and range resolutions to 10m; using the S-G smoothing filter algorithm to remove the time series smoothing thermal noise; radiometric correction; using DEM data for terrain correction.

[0040] Optionally, the features of the crops include band features, texture features, and vegetation index features.

[0041] Optionally, calculating the cumulative distance between the multi-source feature phenological curves of various types of crops and the multi-source feature standard phenological curve includes: setting the time series of the multi-source feature phenological curve of the crop as , with a length of , and the th feature is ; setting the time series of the multi-source feature standard phenological curve as , with a length of , and the th feature is ; then: ; ; Let be the base distance between the th feature and the th feature. Let be the distance matrix between the th feature and the th feature, then: Let the time weight factor be , be the gain factor, be the distance factor. The distance factor is the difference in the time series lengths between the th feature and the th feature. , be the set ratio of the sequence time span, then: ; ; Let be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, be the node corresponding to the th feature and the th feature, and at the node, the cumulative distance is . Find the shortest cumulative distance. The previous node is , and the minimum value between them, then: ; Let the shortest cumulative distance path be L, then: .

[0042] Optionally, when calculating the cumulative distance between the multi-source feature phenological curves and the multi-source feature standard phenological curves of various types of crops, let the cumulative distance of the th crop be , be the The weight of the cumulative distance of each crop, and the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region is For the th crop, the magnitude of the cumulative distance weight is determined by the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological region, then: ; Let be the sum of the cumulative distances between the standard phenological curves in the th phenological region, be the weight of the cumulative distance of the th crop, then: .

[0043] Optionally, determine the crop type of the sample point to be detected according to the magnitude of the cumulative distance and the threshold interval to obtain a preliminary classification sample, and resample the time series of the point to be detected to be consistent with the crop standard time series.

[0044] Optionally, calculate the Manhattan distance between each preliminary classification sample and the corresponding labeled crop. If the Manhattan distance is greater than the set threshold, retain the preliminary classification sample; otherwise, discard the preliminary classification sample. Let the Manhattan distance be , be the eigenvalue of the time series of the point to be detected at time , be the eigenvalue of the crop standard time series at time , be the number of data in the time series, then: .

[0045] Optionally, resample the time series of the point to be detected to be consistent with the crop standard time series before calculating the Manhattan distance between each preliminary classification sample and the corresponding labeled crop.

[0046] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The peanut sample automatic generation method based on multi-source TWDTW is characterized by: include: Several phenological zones are established according to the planting structure of various types of crops in the region, and standard phenological curves corresponding to various characteristics of crops in each phenological zone are established based on the field survey sample point data and pre-processed remote sensing image time series data; The standard phenological curves with different characteristics in the same area and the same time period are spliced ​​into a multi-source characteristic standard phenological curve; Obtain the land cover data of the points to be tested, randomly generate sample points in the cultivated land plot, calculate the time series data of the sample points to be tested, and obtain the multi-source characteristic phenological curves of the sample points to be tested; Calculate the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve, and determine the threshold interval corresponding to the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve; Calculate the cumulative distance between the multi-source characteristic phenological curve of the sample point to be detected and the multi-source characteristic standard phenological curve; The crop type of the sample point to be detected is determined according to the size of the cumulative distance and the threshold interval, and the sample data of the crop type is obtained.

2. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1, characterized in that: Remote sensing image time series data includes SAR data, multispectral remote sensing data and land cover data.

3. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1, characterized in that: The preprocessing of crop remote sensing image time series data includes: filtering the remote sensing image time series data, converting the coordinate system to the world geodetic coordinate system, and resampling the image azimuth and distance resolution to 10m; using the SG smoothing filter algorithm to smooth the time series and remove thermal noise; radiation correction; and terrain correction using DEM data.

4. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1, characterized in that: Crop characteristics include band characteristics, texture characteristics and vegetation index characteristics.

5. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1 is characterized in that: Calculate the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve, including: Assume that the time series of the multi-source characteristic phenological curve of the crop is , the length is , No. The feature is ; Assume that the time series of the multi-source characteristic standard phenological curve is , the length is , No. The feature is ;but: ； ； set up For the Features and The base distance of the features, For the Features and The distance matrix of features is: ； Assume the time weight factor is , is the gain factor, is the distance factor, the distance factor is Features and The difference in the time series length of the features, , is the set ratio of the sequence time span, then: ； ； set up For the Features and The node corresponding to the feature, For the Features and The node corresponding to each feature, For the Features and The node corresponding to the feature, For the Features and The node corresponding to the feature, and exist The cumulative distance at the node is , find the shortest cumulative distance, the previous node is , and The minimum value between , then: ； Assume the minimum cumulative distance path is L, then: 。 6. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1, characterized in that: When calculating the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve, assume that The cumulative distance of crops is , For the The weight of the cumulative distance of each crop, the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological zone is , No. The cumulative distance weight of each crop is determined by the sum of the total distances of the standard phenological curves of each crop in the corresponding phenological zone, then: ； set up For the The cumulative distance between the standard phenological curves in each phenological zone is For the The weight of the cumulative distance of crops is: 。 7. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 1, characterized in that: The crop type of the sample point to be detected is determined according to the size of the cumulative distance and the threshold interval, and the initial sample is obtained. The time series of the point to be detected is resampled to be consistent with the standard time series of the crop.

8. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 5 is characterized in that: Calculate the Manhattan distance between each initial sample and the corresponding labeled crop. If the Manhattan distance is greater than the set threshold, the initial sample is retained, otherwise the initial sample is discarded. Let the Manhattan distance be , is the time series of the points to be detected at time The eigenvalue at For crop standard time series in time The eigenvalue at is the number of data in the time series, then: 。 9. The method for automatically generating peanut samples based on multi-source TWDTW according to claim 8, characterized in that: Before calculating the Manhattan distance between each initial sample and the corresponding labeled crop, the time series of the point to be detected is resampled to be consistent with the standard time series of the crop.

10. The automatic generation system of peanut samples based on multi-source TWDTW is characterized by: It includes a standard phenological curve establishment unit, a splicing unit, a multi-source characteristic phenological curve generation unit, a first data processing unit, a second data processing unit and an output unit; The standard phenological curve establishment unit is used to establish several phenological zones according to the planting structures of various types of crops in the region, and to establish the standard phenological curves corresponding to various characteristics of crops in each phenological zone according to the field survey sample point data and the pre-processed remote sensing image time series data; The splicing unit is used to splice the standard phenological curves with different characteristics in the same area and the same time period into a multi-source characteristic standard phenological curve; A multi-source characteristic phenological curve generating unit is used to obtain land cover data of the points to be detected, randomly generate sample points in the cultivated land plot, calculate the time series data of the sample points to be detected, and obtain the multi-source characteristic phenological curve of the sample points to be detected; The first data processing unit is used to calculate the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve, and determine the threshold interval corresponding to the cumulative distance between the multi-source characteristic phenological curve of each type of crop and the multi-source characteristic standard phenological curve; The second data processing unit is used to calculate the cumulative distance between the multi-source characteristic phenological curve of the sample point to be detected and the multi-source characteristic standard phenological curve; The output unit is used to determine the crop type of the sample point to be detected according to the size of the cumulative distance and the threshold interval, and obtain sample data of the crop type.

Images