Potato late blight area early warning method, device, equipment and storage medium

Abstract

The invention provides a potato late blight area early warning method, device and equipment and a storage medium, and the method comprises the steps: constructing a meteorological data time sequence of a target area based on the meteorological data of the target area, and extracting a potato planting area of the target area based on the satellite remote sensing data of the target area; based on the potato planting area of the target area, calculating daily potato coverage data of the target area; based on the meteorological data time sequence, a CARAH model and the daily potato coverage data, determining an early warning level of each pixel in the target area; the early warning level is used for carrying out early warning on the potato late blight condition in the target area. And early warning can be carried out on the late blight risk condition of the large-area potato vegetation more accurately and comprehensively.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural prevention and control, and in particular to a potato late blight regional early warning method, device, equipment and storage medium. Background Art

[0002] Potato late blight is a common and devastating disease that seriously affects potato yield and quality. It is often called the "cancer" of potatoes. There are two keys to late blight prevention and control. The first is seed potato prevention and control, which includes selecting excellent disease-resistant and insect-resistant varieties and high-quality virus-free seed potatoes, and using seed potato cutter disinfection technology and seed potato treatment (dressing seeds with pesticides) technology. The second is prevention and control during the potato growth period. The most important thing is to find the key period for prevention and control, that is, the spread stage of the central diseased plants. The occurrence of this stage is closely related to meteorological conditions.

[0003] Traditional potato late blight monitoring mainly relies on manual inspections and field surveys, which requires a lot of time and human resources, and is easily interfered by subjective factors, resulting in the accuracy of monitoring being affected. Moreover, it can only provide limited spatial coverage, making it difficult to achieve comprehensive monitoring of large areas, and it is easy to miss or misreport the disease situation. Early warning information is often delayed, making it difficult to provide timely response measures. The CARAH model is a general model for potato late blight. Although this model can be used to draw a potato late blight infection curve through temperature and humidity data obtained from potato field meteorological stations, it can achieve field late blight early warning, but the CARAH model only uses the meteorological data of the station, and the degree of early warning outside the meteorological station is limited. Summary of the invention

[0004] The present invention provides a potato late blight regional early warning method, device, equipment and storage medium, which are used to solve the defects in the prior art that potato late blight monitoring is time-consuming and labor-intensive, it is difficult to cover a large area for comprehensive monitoring, and the CARAH model cannot fully consider the comprehensive information early warning degree of multi-source data, so as to achieve more accurate and comprehensive early warning of the late blight risk status of large areas of potato vegetation.

[0005] The present invention provides a potato late blight regional early warning method, comprising the following steps: Based on the meteorological data of the target area, constructing a time series of the meteorological data of the target area, and based on the satellite remote sensing data of the target area, extracting a potato planting area of ​​the target area; Calculating daily potato coverage data of the target area based on the potato planting area of ​​the target area; Based on the meteorological data time series, the CARAH model and the daily potato coverage data, the warning level of each pixel in the target area is determined; the warning level is used to warn of the potato late blight condition in the target area.

[0006] According to a potato late blight regional early warning method provided by the present invention, based on the meteorological data time series, the CARAH model and the daily potato coverage data, the early warning level of each pixel in the target area is determined, specifically comprising: Calculate the first score of each pixel in the target area according to the meteorological data time series by the Conce scoring method of the CARAH model; Determining a second score of any pixel in the target area according to the daily potato coverage data; Based on the first score and the second score of each pixel in the target area, the warning level of each pixel in the target area is determined.

[0007] According to a potato late blight regional early warning method provided by the present invention, based on the satellite remote sensing data of the target area, the potato planting area of ​​the target area is extracted, which specifically includes: constructing an initial time series image based on satellite remote sensing data of the target area; Based on the initial time series images, a normalized difference vegetation index NDVI image is calculated; Adding the NDVI image to the time series image, reconstructing the time series image, and obtaining a reconstructed time series image; Performing SG filtering on the NDVI image sequence of the reconstructed time series image to smooth the reconstructed time series image; Based on the smoothed reconstructed time series images, the gray-level co-occurrence matrix is ​​calculated to extract texture features; The texture features and ground survey point data are input into a random forest classifier to extract the potato planting area in the target area.

[0008] According to a potato late blight regional early warning method provided by the present invention, the daily potato coverage data of the target area is calculated based on the following formula:

[0009] in, is the normalized difference vegetation index; It is the NDVI value of bare soil or non-vegetated areas in the smoothed reconstructed time series image, which is used to represent areas with no vegetation cover at all; It is the NDVI value of the completely vegetation-covered area in the smoothed reconstructed time series image, which is used to indicate the area completely covered by vegetation.

[0010] According to a potato late blight regional early warning method provided by the present invention, initial time series images are constructed based on satellite remote sensing data of the target area, specifically comprising: Preprocessing the satellite remote sensing data of the target area; Construct initial time series images based on preprocessed satellite remote sensing data; The preprocessing process includes at least one or more of radiation calibration, atmospheric correction and cloud removal algorithm processing.

[0011] According to a potato late blight regional early warning method provided by the present invention, based on the meteorological data of the target area, a meteorological data time series of the target area is constructed, which specifically includes: The meteorological data of the target area are spatially interpolated using the empirical Bayesian kriging method to obtain an interpolation dataset of the meteorological data of the target area; Based on the interpolation data set of meteorological data of the target area, a time series of meteorological data of the target area is constructed.

[0012] The present invention also provides a potato late blight regional early warning device, comprising the following modules: A data extraction module, for constructing a time series of meteorological data of a target area based on the meteorological data of the target area, and for extracting a potato planting area of ​​the target area based on the satellite remote sensing data of the target area; A coverage calculation module, used for calculating daily potato coverage data of the target area based on the potato planting area of ​​the target area; The level determination module is used to determine the warning level of each pixel in the target area based on the meteorological data time series, the CARAH model and the daily potato coverage data; the warning level is used to warn of the potato late blight condition in the target area.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned potato late blight regional early warning methods is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for early warning of potato late blight region as described above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for early warning of potato late blight as described above is implemented.

[0016] The potato late blight regional early warning method, device, equipment and storage medium provided by the present invention monitor meteorological elements and the health status of potato vegetation according to meteorological data and remote sensing data, calculate potato coverage data, and comprehensively analyze the correlation between meteorological data and coverage data and the conditions for the occurrence of late blight, thereby giving early warning of potential late blight risks. It can more accurately and comprehensively warn of the late blight risk status of large areas of potato vegetation, help farmers take timely prevention and control measures, and thus reduce the loss of crops to diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are 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.

[0018] Figure 1 The present invention is a schematic flow chart of the potato late blight regional early warning method.

[0019] Figure 2 This is a schematic diagram of the calculation process of the Conce score index for late blight regional early warning provided by the present invention.

[0020] Figure 3 This is a schematic diagram of the potato coverage FVC extraction process provided by the present invention.

[0021] Figure 4 A flow chart of a regional early warning device and method for potato late blight coupled with remote sensing and meteorological factors provided by the present invention.

[0022] Figure 5 This is a schematic diagram of the distribution of meteorological stations in the research area provided by the present invention.

[0023] Figure 6 The interpolation map of relative humidity on different dates in the study area provided by the present invention.

[0024] Figure 7 This is an interpolation map of the temperature on different dates in the study area provided by the present invention.

[0025] Figure 8 This is a distribution map of ground survey points in the research area provided by the present invention.

[0026] Fig. 9This is a schematic diagram of the potato planting area extraction results provided by the present invention.

[0027] Fig.10 This is a schematic diagram of the FVC monitoring results of vegetation coverage in the study area at different periods provided by the present invention.

[0028] Fig.11 This is a schematic diagram of the FVC monitoring results of potato coverage in the study area at different periods provided by the present invention.

[0029] Fig.12 This is the early warning information map of the study area at different periods provided by the present invention.

[0030] Fig.13 A schematic diagram of the infection cycle process obtained by using the Couce score provided by the present invention.

[0031] Fig.14 This is a schematic diagram of the land risk warning results provided by the present invention.

[0032] Fig.15 The diagram is a schematic structural diagram of a potato late blight regional early warning device provided by the present invention.

[0033] Fig.16 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0035] Figure 1 FIG. 1 is a flow chart of a regional early warning method for potato late blight provided by the present invention, such as Figure 1 As shown, the method comprises the following steps: Step 100: construct a time series of meteorological data of the target area based on the meteorological data of the target area, and extract a potato planting area of ​​the target area based on the satellite remote sensing data of the target area.

[0036] Step 101: Calculate daily potato coverage data of the target area based on the potato planting area of ​​the target area.

[0037] Step 102: Determine the warning level of each pixel in the target area based on the meteorological data time series, the CARAH model and the daily potato coverage data; the warning level is used to warn of the potato late blight condition in the target area.

[0038] Specifically, the target area in the embodiment of the present invention refers to a potato planting area where the development of late blight needs to be monitored. In the embodiment of the present invention, meteorological data and satellite remote sensing data of the target area can be firstly obtained respectively.

[0039] Common meteorological data include temperature, relative humidity, rainfall, wind speed, light, etc. The embodiment of the present invention does not limit the source of meteorological data, and meteorological data can be collected from multiple sources. For example, the meteorological data used in the embodiment of the present invention can come from the hourly data set of basic ground observation data in China provided by the National Data Meteorological Network. For another example, the meteorological data used in the embodiment of the present invention can also include hourly meteorological data from the late blight monitoring station, and the relevant meteorological data is updated daily.

[0040] After obtaining the meteorological data of the target area, a meteorological data time series of the target area can be constructed based on the meteorological data of the target area.

[0041] According to a potato late blight regional early warning method provided by the present invention, based on the meteorological data of the target area, a meteorological data time series of the target area is constructed, which specifically includes: The meteorological data of the target area is spatially interpolated using the empirical Bayesian kriging method to obtain an interpolation dataset of meteorological data of the target area; Based on the interpolation dataset of meteorological data in the target area, a time series of meteorological data in the target area is constructed.

[0042] Specifically, meteorological stations may be sparsely distributed and unable to cover the entire target area. The meteorological data of the target area obtained from various sources may also have missing data. Therefore, the empirical Bayesian Kriging (EBK) method can be used to perform spatial interpolation processing on the meteorological data of the target area. By analyzing the spatial correlation between the meteorological station data, the interpolation dataset of the meteorological data of the entire target area can be obtained.

[0043] After obtaining the interpolated meteorological data set of the entire target area, a corresponding meteorological data time series can be created for each pixel in the target area based on its coordinates and time. That is, for each pixel, the value of the meteorological variable (temperature, humidity, etc.) is recorded at each time point.

[0044] Optionally, the time series can be recorded at different time scales such as hours, days, months, etc., depending on the application requirements. For example, the embodiment of the present invention is applied to agriculture, so hourly meteorological data can better track the impact of microclimate on crop growth, and hourly meteorological data can be used to construct hourly meteorological data time series.

[0045] Optionally, the satellite remote sensing data used in the embodiment of the present invention may be Sentinel-2 L1C data, which is derived from the European Space Agency (ESA) Copernicus Open Access Hub, and the data may provide a multi-band, high-resolution potato growth monitoring map of the study area. The Sentinel-2A / 2B satellite group has a high spatial resolution, typically 10 meters, and a revisit period of typically 2 to 5 days, and may be used in the embodiment of the present invention to monitor potato planting plot extraction, potato growth, and coverage monitoring.

[0046] After obtaining the satellite remote sensing data of the target area, the potato planting area of ​​the target area can be extracted based on the satellite remote sensing data of the target area. Ground sample points of the potato planting area in the target area can be collected as reference data for classification, and the existing machine learning algorithm can be used to supervise the learning and classification of the satellite remote sensing images of the target area, so as to extract the potato planting area of ​​the target area.

[0047] After extracting the potato planting area of ​​the target area, the normalized difference vegetation index (NDVI) or enhanced vegetation index (EVI) can be used to calculate the daily potato coverage data of the target area, that is, the fractional vegetation cover (FVC) data, based on the potato planting area.

[0048] After obtaining the meteorological data time series and daily potato coverage data, the warning level of each pixel in the target area can be determined according to the meteorological data time series, the CARAH model and the daily potato coverage data, and then the potato late blight situation in the target area can be warned according to different warning levels.

[0049] The CARAH model is a model developed in Belgium specifically for early warning of potato late blight. The model evaluates the potential risk of potato late blight by analyzing meteorological station data, such as temperature and relative humidity. It determines the spread conditions of potato late blight based on the growth environment requirements of the late blight pathogen, especially changes in temperature and humidity.

[0050] In the CARAH model, the key analysis stage is to identify and study the period of time when the relative humidity exceeds 90%, which is usually defined as the "wet period". During this wet period, the temperature data can be averaged hour by hour to obtain the average temperature of the wet period. The potential infection degree of late blight is then calculated according to the Guntz-Divoux model (Table 1). Between 7-27℃, the longer the wet period, the more severe the expected disease. Rainfall can increase air humidity, hinder the use of fungicides and reduce the effectiveness of pesticides.

[0051] The assessment of the degree of late blight infection depends on the duration of the wet period and the average temperature during this period. The study found that the longer the relative humidity lasts above 90% and the higher the average temperature, the more serious the late blight infection. This correlation is quantified in Table 1. However, it is worth noting that when the temperature is below 7°C or above 30°C, the growth of late blight fungi is inhibited, and even if the relative humidity meets the growth conditions, it will not lead to the occurrence of late blight.

[0052] Different thresholds for infection levels vary at different average temperatures during the wet period. For example, according to Table 1, when the average temperature during the wet period is 7°C, it takes at least 16 hours and 30 minutes for mild infection to occur, while when the average temperature is 15°C, a wet period of 10 hours and 45 minutes is sufficient to cause mild infection. This shows that, regardless of the average temperature, late blight is likely to occur once the wet period exceeds a certain length of time, which is usually more than 8 hours. In addition, if the interruption during the wet period does not exceed 3 hours, it is regarded as a continuous wet period; if the interruption exceeds 4 hours, it is regarded as two separate wet periods.

[0053] Once the conditions described in Table 1 are met, if late blight spores are present in the area, the infection process will begin, independent of the relative humidity outside. This is because the relative humidity inside the leaves can be maintained at 100%, providing an ideal growth environment for late blight.

[0054] Table 1. Relationship between the potential infection level of late blight and the average temperature during the wet period according to the Guntz-Divoux model

[0055] Since the spread and infection of late blight spores require good humidity and host conditions, high potato NDVI values ​​usually indicate vigorous vegetation growth, and higher vegetation cover (FVC) indicates that there are more potato tissues in the environment. This means that when the relative humidity and temperature of the air are high, high vegetation cover will further reduce the air permeability of the lower layer of vegetation, increase its moisture level, and increase the risk of late blight spread.

[0056] In summary, the risk of potato late blight can be determined by the meteorological data time series and the CARAH model. On this basis, the embodiment of the present invention combines the daily potato coverage data to more comprehensively determine the risk of potato late blight.

[0057] The potato late blight regional early warning method provided by the present invention monitors meteorological elements and the health status of potato vegetation according to meteorological data and remote sensing data, calculates potato coverage data, and comprehensively analyzes the correlation between meteorological data and coverage data and the conditions for the occurrence of late blight, thereby giving early warning of potential late blight risks. It can more accurately and comprehensively warn of the late blight risk status of large areas of potato vegetation, help farmers take timely prevention and control measures, and thus reduce the loss of crops caused by diseases.

[0058] According to a potato late blight regional early warning method provided by the present invention, based on the meteorological data time series, the CARAH model and the daily potato coverage data, the early warning level of each pixel in the target area is determined, specifically including: The first score of each pixel in the target area is calculated based on the meteorological data time series by the Conce score method of the CARAH model; Determine the second score of any pixel in the target area based on the daily potato coverage data; Based on the first score and the second score of each pixel in the target area, the warning level of each pixel in the target area is determined.

[0059] Specifically, the first score of each pixel in the target area can be calculated based on the meteorological data time series by using the Conce score method of the CARAH model.

[0060] The following is an introduction to the existing Concept scoring method of the CARAH model.

[0061] Based on the CARAH model and years of observation experience, the embodiment of the present invention aims to explore the connotation and characteristics of the late blight infection cycle. The daily score is calculated according to the Conce method. The specific score can be referred to Table 2. In the CARAH model, each infection cycle needs to obtain a Conce score of 7 to complete the process of spore maturation, spore germination and new spore infection. Once it reaches 7 points, the infection cycle ends. If spraying is not taken at this time, lesions and white mycelium will be seen on the leaves.

[0062] Table 2 Conce score calculation

[0063] It should be noted that the Conce score is generally used to start calculation when the central diseased plant is found or the late blight spores are detected, but in the embodiment of the present invention, it is used to issue an early warning for the pixels in the potato planting area that meet the infection conditions.

[0064] After the infection cycle begins, as long as the average temperature of each day is obtained, a score can be obtained according to the data provided in Table 2, and then the scores obtained each day are accumulated. If 7 points are obtained, it means that an infection cycle has ended and a new infection cycle is about to begin. As can be seen from the table, at 16-20℃, the score is very high, because this temperature range is the most suitable temperature range for the growth of late blight. If the temperature is low, the more days it takes to reach 7 points, the more time it may take to complete an infection cycle; when the temperature is high, it may only take 4-5 days to complete an infection cycle.

[0065] In the embodiment of the present invention, the influence of the daily potato coverage data on the risk of late blight is taken into consideration. Therefore, it is also necessary to determine the second score of any pixel in the target area according to the daily potato coverage data.

[0066] In the embodiment of the present invention, if FVC>0.60, the second score is 0.5, otherwise, the second score is 0.

[0067] Then, the warning level of each pixel in the target area may be determined according to the first score and the second score of each pixel in the target area.

[0068] It is understandable that based on the original Conce score, for areas that meet the infection conditions, the NDVI curve fitted by the time series can be used to judge that if FVC>0.60, the Conce score will increase by 0.5. However, it should be noted that after research, in this case, the threshold of the Conce score should be greater than 8 points to complete an infection cycle.

[0069] The process of determining the warning level of each pixel in the target area of ​​the present invention is described in detail below through specific embodiments.

[0070] As mentioned above, when conducting an early warning of potential damage from late blight in a region, the key analysis stage is to identify and study the time period when the relative humidity exceeds 90%, which is usually defined as the "wet period". During this wet period, the temperature data can be averaged hour by hour to obtain the average temperature of the wet period.

[0071] Figure 2 The schematic diagram of the calculation process of the Conce score index for the late blight regional early warning provided by the present invention is as follows: Figure 2 As shown in the figure, the Concept score is based on pixel-by-pixel processing of hourly relative humidity, temperature and coverage data images of the study area.

[0072] The embodiment of the present invention can obtain hourly meteorological data from different meteorological stations starting from the potato seedling stage, perform spatial interpolation on the obtained relative humidity and temperature data, and first determine whether the relative humidity value of each pixel in the study area is greater than 90%. Once a pixel with humidity greater than 90% is found, the duration of the pixel humidity greater than 90% (the duration of the wet period) and the average temperature during the wet period are calculated. The invention obtains the Conce score (i.e., the first score) of each potato pixel based on the relationship between the potential infection degree of late blight and the average temperature during the wet period according to the Guntz-Divoux model (Table 1). Among them, for pixels with humidity greater than 90%, the size of their vegetation coverage FVC can be further determined. When the pixel FVC value is greater than 0.6, the pixel Conce score increases by 0.5 (i.e., the second score is 0.5).

[0073] The embodiment of the present invention calculates the Conce score (i.e., the sum of the first score and the second score) for the pixels that meet the infection conditions. According to the classification standard in Table 1, the warning level can be divided into 4 levels, namely, mild, medium, severe, and extremely severe. In the embodiment of the present invention, the warning level will be increased by 1 level every time the Conce score reaches 7 points. An early warning can be issued when the mild potential infection standard is reached, and the warning level is 1. The warning levels for medium, severe, and extremely severe potential infection standards are 2, 3, and 4. It has been found through research that when the warning level is 2, it is sufficient to investigate the late blight center diseased plants on the ground, and spraying late blight is very effective at this time.

[0074] In some implementations, the Conce score index can be calculated for each pixel's hourly meteorological data and pixel coverage FVC information, and ultimately the late blight area infection information of the study area at different time periods can be obtained. On this basis, the potato field boundary information can be combined to assign values ​​to the fields according to the pixel value with the highest Conce score within the field, ultimately achieving field-scale potato late blight early warning.

[0075] In some embodiments, the potato coverage FVC calculation in the embodiments of the present invention is completed in GEE, and the potato late blight infection process judgment, potato late blight Conce score calculation, potato late blight area and plot warning can all be calculated pixel by pixel in the study area using Python.

[0076] According to a potato late blight regional early warning method provided by the present invention, based on satellite remote sensing data of the target area, the potato planting area of ​​the target area is extracted, which specifically includes: Construct initial time series images based on satellite remote sensing data of the target area; Based on the initial time series images, calculate the Normalized Difference Vegetation Index (NDVI) image; The NDVI image is added to the time series image, and the time series image is reconstructed to obtain the reconstructed time series image; Perform SG filtering on the NDVI image sequence of the reconstructed time series image to smoothly reconstruct the time series image; Based on the smoothed reconstructed time series images, the gray-level co-occurrence matrix is ​​calculated to extract texture features; The texture features and ground survey point data were input into the random forest classifier to extract the potato planting area in the target area.

[0077] Specifically, the embodiment of the present invention can use the random forest algorithm to extract the potato planting area based on potato ground sample point data.

[0078] Since the potato planting process takes time and crop growth conditions change over time, we can first construct initial time series images based on satellite remote sensing data of the target area to capture important changes in the crop growth cycle.

[0079] Then, the NDVI image can be calculated for each time node image and added to the time series. In some implementations, in order to reduce cloud interference and make the time series curve smoother, the time series can be reconstructed by taking the average of ten days of images to obtain a reconstructed time series image.

[0080] After obtaining the reconstructed time series image, the NDVI image sequence of the time series image can be processed by SG filtering (Savitzky-Golay filter) to smooth the time series curve and obtain the smoothed reconstructed time series image.

[0081] Then, the gray-level co-occurrence matrix (GLCM) can be calculated based on the smoothed reconstructed time series images to extract texture features. This can then be combined with ground survey point data, such as the digital elevation model (DEM) data provided by NASA SRTM Digital Elevation 30m, and input into a random forest classifier to extract potato planting areas.

[0082] Optionally, when the potato planting area of ​​the target area is extracted in the above manner, the daily potato coverage data of the target area can be calculated based on the following formula:

[0083] in, is the normalized difference vegetation index; It is the NDVI value of bare soil or non-vegetated areas in the smoothed reconstructed time series image, which is used to represent areas with no vegetation cover at all; It is the NDVI value of the completely vegetation-covered area in the smoothed reconstructed time series image, which is used to indicate the area completely covered by vegetation.

[0084] Figure 3 The schematic diagram of the potato coverage FVC extraction process provided by the present invention is as follows: Figure 3 As shown, in some embodiments, before the satellite remote sensing data (such as Sentinel-2 data) is updated, the daily FVC and NDVI are calculated by gradually updating the actual monitoring data from the potato planting period of the year throughout the potato growing period, and using a linear interpolation function to predict the daily FVC and NDVI until the satellite remote sensing data is updated. It can be understood that although FVC and NDVI are updated every 2-5 days following the Sentinel-2 data, they are still sufficient to indicate the conditions for the occurrence of late blight.

[0085] According to a potato late blight regional early warning method provided by the present invention, initial time series images are constructed based on satellite remote sensing data of the target area, specifically comprising: Preprocessing of satellite remote sensing data of target area; Construct initial time series images based on preprocessed satellite remote sensing data; The preprocessing process includes at least one or more of radiation calibration, atmospheric correction and cloud removal algorithm processing.

[0086] Specifically, after obtaining the satellite remote sensing data of the target area, the satellite remote sensing data of the target area may be preprocessed first, and then an initial time series image may be constructed based on the preprocessed satellite remote sensing data.

[0087] The preprocessing process may include one or more of radiation calibration, atmospheric correction and cloud removal algorithm processing.

[0088] Among them, the radiation calibration and atmospheric correction process can be implemented through the SNAP (Sentinel Application Platform) toolbox, thereby processing the L1C data into L2A data.

[0089] Since cloud interference in satellite remote sensing data will affect the recognition accuracy, the cloud removal algorithm (such as threshold method based on blue light band, NDVI, Normalized Difference Snow Index (NDSI) etc.) can be used to remove the cloud influence and ensure the accuracy of the data.

[0090] In some implementations, the de-clouding algorithm may be described as follows: (1) Normalize the pixel values ​​of the blue light band (B2) within the range of [0.1, 0.5]; (2) Normalize the pixel values ​​of the green band (B1) within the range [0.1, 0.3]; (3) The pixel values ​​after normalization processing of the green band (B1) and the near-infrared band (B9) are within the range of [0.15, 0.2]; (4) Normalize the sum of the pixel values ​​of the blue, green, and red bands (B2, B3, and B4) within the range [0.2, 0.8]; (5) Normalization: the part where the Normalized Difference Moisture Index (NDMI) value is within the range of [-0.3, 0.3]; (6) Normalize the NDSI values ​​within the range of [0.9, 0.4].

[0091] Take the minimum value of all the above values ​​and compare it with the cloud threshold of 0.12. If it is greater than the threshold, it is judged as a cloud.

[0092] The formula used is as follows:

[0093] Wherein, NDVI is the normalized vegetation index; EVI is the enhanced vegetation index; NDMI is the normalized difference moisture index; NDSI is the normalized difference snow index. is the reflectivity in the near-infrared band; is the reflectivity of the red band; is the reflectivity of the blue band; is the reflectivity in the infrared band; is the reflectivity of the green band; is the reflectivity in the shortwave infrared band.

[0094] In some implementations, the satellite remote sensing data processing, NDVI, EVI and other parameter calculations, satellite remote sensing data image time series construction, time series image SG filtering, and NDVI image data grayscale co-occurrence matrix generation and texture feature extraction in the embodiments of the present invention are all completed in the GEE (Google Earth Engine) platform.

[0095] The key points and practical effects of the potato late blight regional early warning method provided by the present invention are further explained below through embodiments in specific application scenarios.

[0096] Figure 4The flowchart of the potato late blight regional early warning device and method coupled with remote sensing and meteorological factors provided by the present invention is as follows: Figure 4 As shown, this embodiment integrates multi-source data, including meteorological data, remote sensing data, etc. These data come from different platforms and sensors and need to be effectively integrated and processed. Advanced data processing and analysis technologies, including remote sensing data processing, image processing, data mining, and machine learning, are used to extract key information about late blight from a large amount of data.

[0097] This embodiment uses temperature and relative humidity as the key meteorological elements for late blight prediction. It starts to collect meteorological data synchronously after potato sowing in the study area. Based on the data of each monitoring station, Python scripts are written in ArcGIS for batch processing. The hourly meteorological element data of each station are spatially interpolated using the empirical Bayesian Kriging method to obtain the hourly meteorological data interpolation data set during the potato growing period in the study area, which is used to construct the meteorological data time series of the study area. Secondly, this embodiment completes the Sentinel-2 data processing, NDVI, EVI and other parameter calculations of the study area in the GEE platform, constructs the Sentinel-2 image time series data, filters based on the SG algorithm, and combines the grayscale co-occurrence matrix and texture features of the NDVI image data of the study area to realize the extraction of the potato planting area in the study area in the current season. Combined with the time-series Sentinel-2 image, the potato coverage FVC of the study area is extracted. Finally, with the support of regional hourly temperature and humidity data, this embodiment combines the humidity and temperature information in the Guntz-Divoux model with the potato coverage FVC remote sensing monitoring information, uses Python programming to judge the potato late blight infection process pixel by pixel in the study area, and calculates the potato late blight Conce score, ultimately achieving potato late blight regional and plot early warning.

[0098] This embodiment takes into account various meteorological conditions based on the CARAH model, analyzes the temporal and spatial changes of these factors, and establishes a time series model to predict potential late blight risks. At the same time, combined with remote sensing information, the regional late blight regional early warning method and Conce score calculation are improved, which helps to further optimize the large-scale late blight prevention and control strategy.

[0099] The following describes each part of the process of this embodiment.

[0100] 1. Data acquisition and preprocessing

[0101] 1.1 Meteorological data collection and preprocessing

[0102] This embodiment takes Keshan County and Nehe City of Heilongjiang Province as the research area. Through the implementation of this embodiment, the potato late blight early warning map and late blight spore infection cycle map of the experimental area in 2023 are finally obtained.

[0103] This embodiment uses data provided by meteorological stations and potato late blight monitoring stations set up in the study area as a regional early warning for potato late blight in 2023 based on the CARAH model.

[0104] Figure 5 Schematic diagram of the distribution of meteorological stations in the research area provided by the present invention, wherein the meteorological data uses the hourly ground basic meteorological observation data from June 1 to September 30, 2023 provided by the China Meteorological Data Network, in addition to the following Figure 5 The meteorological data provided by the 2022-2023 potato late blight monitoring station is shown in the figure. The meteorological factors used include temperature (℃) and relative humidity. The meteorological data from different sources are fused and interpolated by the empirical Bayesian kriging method to obtain the meteorological data interpolation map of the study area. Figure 6 The interpolation diagram of relative humidity on different dates in the research area provided by the present invention is: Figure 7 This is an interpolation map of the temperature on different dates in the study area provided by the present invention.

[0105] 1.2 Remote sensing extraction of potato planting areas

[0106] The implementation case of this embodiment uses the Sentinel-2L1C data from 2022 to 2023, uses the SNAP toolbox for radiation calibration, atmospheric correction and other preprocessing, and performs cloud removal through the method in the invention scheme. The classification method uses time-weighted random forest classification to construct Sentinel-2 time series images, calculate NDVI and gray-level co-occurrence matrix (GLCM) features, and add them to the image collection. The image needs to be synthesized for 5 days to ensure its temporal regularity. Then use SG filtering to smooth the time series image values. Potato planting areas are extracted based on the smoothed time series data and ground survey point data.

[0107] In order to accurately extract potato objects, this embodiment of the invention conducted a survey on the types of crops in the research area. Figure 8 The ground survey point distribution map of the research area provided by the present invention, Fig. 9 This is a schematic diagram of the potato planting area extraction results provided by the present invention.

[0108] 1.3 Potato FVC calculation

[0109] Fig.10This is a schematic diagram of the FVC monitoring results of vegetation coverage in the study area at different periods provided by the present invention. Based on the obtained Sentinel2 image in 2023, the NDVI value of the study area is calculated to obtain the spatial distribution map of the FVC vegetation coverage in the study area. The results are as follows Fig.10 shown.

[0110] Fig.11 The schematic diagram of the potato coverage FVC monitoring results at different periods in the study area provided by the present invention is to obtain the potato field FVC value based on the obtained potato planting area classification results in the study area. The results are as follows Fig.11 shown.

[0111] 2. Early warning of potato late blight infection in regions

[0112] The pre-processed meteorological image data and coverage image data are processed pixel by pixel. The process is as follows: Figure 3 As shown, the wet period time (the number of hours when the humidity is greater than 90%) and the average temperature during the wet period are calculated for each pixel. For each pixel in the study area, first determine whether its relative humidity value is greater than 90%. Once a pixel with humidity greater than 90% is found, the duration of the humidity greater than 90% (the duration of the wet period) and the average temperature during the wet period are calculated. The invention obtains the Conce score of each potato pixel based on the relationship between the potential infection degree of late blight and the average temperature during the wet period according to the Guntz-Divoux model (Table 1). For pixels with humidity greater than 90%, the size of their vegetation coverage FVC is further determined. When the pixel FVC value is greater than 0.6, the pixel Conce score increases by 0.5.

[0113] The above algorithm is applied to the hourly meteorological data of each pixel. According to the classification standards in Table 1, the warning level is divided into 5 levels. If the warning level reaches level 2, it is possible to find the central diseased plants in the field. Fig.12 The early warning information map of the study area provided by the present invention at different times includes Nehe City and Keshan County.

[0114] When there is a risk level, the Conce score can be calculated. According to the method in Table 2, the infection process is calculated by combining temperature and FVC for each hour of risk area. The infection process will be completed within 10 days. If there are many green leaves during the growth period and the temperature is suitable, the infection can be completed in about 4 days. The next round of infection cycle will start.

[0115] Fig.13 The schematic diagram of the infection cycle process obtained by using the Couce score provided by the present invention is as follows: Fig.13As shown, it shows the infection cycle process obtained by the Conce score of potato late blight in a certain image unit in Keshan County from June 1 to September 20 obtained according to the above method.

[0116] 3. Early warning of potato late blight field infection

[0117] On the basis of the above research, combined with the potato classification results and potato plot boundary information, the Conce score of the plot was extracted according to the boundaries of different plots, and the late blight infection warning information of the plot was issued according to the pixel with the highest Conce score in the plot. Fig.14 A schematic diagram of the land risk warning result provided by the present invention, such as Fig.14 As shown, it shows the late blight risk warning results of this embodiment for different potato plots in Keshan Farm in 2023.

[0118] This embodiment can monitor meteorological elements and the health of potato vegetation based on meteorological data and remote sensing data. By comprehensively analyzing the relevance of these data to the conditions for the occurrence of late blight, early warning of potential late blight risks can be provided to help farmers take timely prevention and control measures, thereby reducing crop losses caused by diseases. Farmers can also manage farmland more effectively, which helps to improve the stability and yield of agricultural production, avoid unnecessary use of pesticides, reduce costs, and reduce environmental impacts. This embodiment provides a scientific basis to help government agencies and decision makers formulate better agricultural policies and management strategies to support sustainable agricultural development and food security.

[0119] The potato late blight regional early warning device provided by the present invention is described below. The potato late blight regional early warning device described below and the potato late blight regional early warning method described above can be referred to each other.

[0120] Fig.15 : is a schematic diagram of the structure of the potato late blight regional early warning device provided by the present invention, such as Fig.15 As shown, the device includes the following modules: The data extraction module 1500 is used to construct a time series of meteorological data of the target area based on the meteorological data of the target area, and to extract a potato planting area of ​​the target area based on the satellite remote sensing data of the target area; A coverage calculation module 1510 is used to calculate daily potato coverage data of a target area based on a potato planting area of ​​the target area; The level determination module 1520 is used to determine the warning level of each pixel in the target area based on the meteorological data time series, the CARAH model and the daily potato coverage data; the warning level is used to warn the potato late blight situation in the target area.

[0121] According to a potato late blight regional early warning device provided by the present invention, based on the meteorological data time series, the CARAH model and the daily potato coverage data, the early warning level of each pixel in the target area is determined, specifically including: The first score of each pixel in the target area is calculated based on the meteorological data time series by the Conce score method of the CARAH model; Determine the second score of any pixel in the target area based on the daily potato coverage data; Based on the first score and the second score of each pixel in the target area, the warning level of each pixel in the target area is determined.

[0122] According to a potato late blight regional early warning device provided by the present invention, based on satellite remote sensing data of the target area, the potato planting area of ​​the target area is extracted, specifically comprising: Construct initial time series images based on satellite remote sensing data of the target area; Based on the initial time series images, calculate the Normalized Difference Vegetation Index (NDVI) image; The NDVI image is added to the time series image, and the time series image is reconstructed to obtain the reconstructed time series image; Perform SG filtering on the NDVI image sequence of the reconstructed time series image to smoothly reconstruct the time series image; Based on the smoothed reconstructed time series images, the gray-level co-occurrence matrix is ​​calculated to extract texture features; The texture features and ground survey point data were input into the random forest classifier to extract the potato planting area in the target area.

[0123] According to a potato late blight regional early warning device provided by the present invention, the daily potato coverage data of the target area is calculated based on the following formula:

[0124] in, is the normalized difference vegetation index; It is the NDVI value of bare soil or non-vegetated areas in the smoothed reconstructed time series image, which is used to represent areas with no vegetation cover at all; It is the NDVI value of the completely vegetation-covered area in the smoothed reconstructed time series image, which is used to indicate the area completely covered by vegetation.

[0125] According to a potato late blight regional early warning device provided by the present invention, an initial time series image is constructed based on satellite remote sensing data of a target area, specifically comprising: Preprocessing of satellite remote sensing data of target area; Construct initial time series images based on preprocessed satellite remote sensing data; The preprocessing process includes at least one or more of radiation calibration, atmospheric correction and cloud removal algorithm processing.

[0126] According to a potato late blight regional early warning device provided by the present invention, based on the meteorological data of the target area, a meteorological data time series of the target area is constructed, which specifically includes: The meteorological data of the target area is spatially interpolated using the empirical Bayesian kriging method to obtain an interpolation dataset of meteorological data of the target area; Based on the interpolation dataset of meteorological data in the target area, a time series of meteorological data in the target area is constructed.

[0127] Fig.16 A schematic diagram of the structure of an electronic device provided by the present invention, such as Fig.16 As shown, the electronic device may include: a processor 1610, a communications interface 1620, a memory 1630 and a communication bus 1640, wherein the processor 1610, the communications interface 1620 and the memory 1630 communicate with each other through the communication bus 1640. The processor 1610 may call the logic instructions in the memory 1630 to execute the potato late blight regional early warning method, which includes: Based on the meteorological data of the target area, a time series of meteorological data of the target area is constructed, and based on the satellite remote sensing data of the target area, a potato planting area of ​​the target area is extracted; Calculate daily potato coverage data for the target area based on the potato planting area in the target area; Based on the meteorological data time series, CARAH model and daily potato coverage data, the warning level of each pixel in the target area is determined; the warning level is used to warn of the potato late blight situation in the target area.

[0128] In addition, the logic instructions in the above-mentioned memory 1630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0129] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the potato late blight regional early warning method provided by the above methods, the method comprising: Based on the meteorological data of the target area, a time series of meteorological data of the target area is constructed, and based on the satellite remote sensing data of the target area, a potato planting area of ​​the target area is extracted; Calculate daily potato coverage data for the target area based on the potato planting area in the target area; Based on the meteorological data time series, CARAH model and daily potato coverage data, the warning level of each pixel in the target area is determined; the warning level is used to warn of the potato late blight situation in the target area.

[0130] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the potato late blight regional early warning method provided by the above methods, the method comprising: Based on the meteorological data of the target area, a time series of meteorological data of the target area is constructed, and based on the satellite remote sensing data of the target area, a potato planting area of ​​the target area is extracted; Calculate daily potato coverage data for the target area based on the potato planting area in the target area; Based on the meteorological data time series, CARAH model and daily potato coverage data, the warning level of each pixel in the target area is determined; the warning level is used to warn of the potato late blight situation in the target area.

[0131] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0132] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A potato late blight regional early warning method, characterized in that: include: Based on the meteorological data of the target area, constructing a time series of the meteorological data of the target area, and based on the satellite remote sensing data of the target area, extracting a potato planting area of ​​the target area; Calculating daily potato coverage data of the target area based on the potato planting area of ​​the target area; Based on the meteorological data time series, the CARAH model and the daily potato coverage data, the warning level of each pixel in the target area is determined; the warning level is used to warn of the potato late blight condition in the target area.

2. The potato late blight regional early warning method according to claim 1, characterized in that: Based on the meteorological data time series, the CARAH model and the daily potato coverage data, the warning level of each pixel in the target area is determined, specifically including: Calculate the first score of each pixel in the target area according to the meteorological data time series by the Conce scoring method of the CARAH model; Determining a second score of any pixel in the target area according to the daily potato coverage data; Based on the first score and the second score of each pixel in the target area, the warning level of each pixel in the target area is determined.

3. The potato late blight regional early warning method according to claim 1, characterized in that: Extracting the potato planting area of ​​the target area based on the satellite remote sensing data of the target area specifically includes: constructing an initial time series image based on satellite remote sensing data of the target area; Based on the initial time series images, a normalized difference vegetation index NDVI image is calculated; Adding the NDVI image to the time series image, reconstructing the time series image, and obtaining a reconstructed time series image; Performing SG filtering on the NDVI image sequence of the reconstructed time series image to smooth the reconstructed time series image; Based on the smoothed reconstructed time series images, the gray-level co-occurrence matrix is ​​calculated to extract texture features; The texture features and ground survey point data are input into a random forest classifier to extract the potato planting area in the target area.

4. The potato late blight regional early warning method according to claim 3, characterized in that: The daily potato coverage data for the target area is calculated based on the following formula: in, is the normalized difference vegetation index; It is the NDVI value of bare soil or non-vegetated areas in the smoothed reconstructed time series image, which is used to represent areas with no vegetation cover at all; It is the NDVI value of the fully vegetation-covered area in the smoothed reconstructed time series image, which is used to indicate the area completely covered by vegetation.

5. The potato late blight regional early warning method according to claim 3, characterized in that: Based on the satellite remote sensing data of the target area, construct an initial time series image, specifically including: Preprocessing the satellite remote sensing data of the target area; Construct initial time series images based on preprocessed satellite remote sensing data; The preprocessing process includes at least one or more of radiation calibration, atmospheric correction and cloud removal algorithm processing.

6. The potato late blight regional early warning method according to claim 1, characterized in that: Based on the meteorological data of the target area, a time series of meteorological data of the target area is constructed, specifically including: The meteorological data of the target area are spatially interpolated using the empirical Bayesian kriging method to obtain an interpolation dataset of the meteorological data of the target area; Based on the interpolation data set of meteorological data of the target area, a time series of meteorological data of the target area is constructed.

7. A potato late blight regional early warning device, characterized in that: include: A data extraction module, for constructing a time series of meteorological data of a target area based on the meteorological data of the target area, and for extracting a potato planting area of ​​the target area based on the satellite remote sensing data of the target area; A coverage calculation module, used for calculating daily potato coverage data of the target area based on the potato planting area of ​​the target area; The level determination module is used to determine the warning level of each pixel in the target area based on the meteorological data time series, the CARAH model and the daily potato coverage data; the warning level is used to warn of the potato late blight condition in the target area.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the potato late blight regional early warning method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the potato late blight regional early warning method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the potato late blight regional early warning method according to any one of claims 1 to 6 is implemented.

Images