A method, system, electronic device and medium for remotely monitoring corn lodging

Abstract

The present invention relates to a method, system, electronic device and medium for remote sensing monitoring of maize lodging, belonging to the field of remote sensing monitoring. The method extracts sentinel data at modeling sample points through field sampling; screens and preprocesses the sentinel data at the modeling sample points to obtain full-time sentinel-1 dual-polarization data and pre-lodging sentinel-2 optical image data, and calculates the pre-lodging normalized difference vegetation index of maize and the difference in the best sensitive parameters during the maize disaster period and the recovery period based on this, so as to divide all modeling sample points into three levels: low vitality, medium vitality and high vitality, and respectively count the mean value of the difference in the best sensitive parameters during the maize disaster period and the recovery period as the statistical threshold at each level to establish a maize lodging monitoring decision tree; determines the maize lodging degree of the monitored maize planting area according to the maize lodging monitoring decision tree. The method of the present invention can monitor the maize lodging situation in a large area and effectively improve the monitoring accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing monitoring, and particularly to a method, system, electronic device and medium for remote sensing monitoring of corn lodging. Background Technique

[0002] The lodging phenomenon of corn during the growing season is an important factor restricting the yield of corn crops. Lodging not only causes serious losses in corn yield, but also large-scale lodging is not conducive to later mechanical harvesting and will increase the harvesting cost. Rapid, accurate and large-scale monitoring of corn lodging can provide timely information feedback for agricultural management, which is of great significance for reducing corn yield losses and economic losses. Traditional crop lodging assessment methods mainly rely on on-site visual rating by experts and manual measurement, which have certain subjectivity, high costs and low timeliness. Monitoring based on remote sensing technology is a non-field and non-contact monitoring method, which avoids a large amount of human consumption and can simply and efficiently obtain the real-time lodging situation of corn. However, the existing lodging monitoring methods using remote sensing technology often have the following deficiencies: on the one hand, the spatio-temporal resolution of remote sensing data and the correlation of spectral band information restrict the quantitative monitoring of remote sensing data; on the other hand, due to the complex farmland conditions, it increases the difficulty of extracting optical information from optical data before and after lodging occurs, and in extreme weather, optical data cannot be obtained in a timely and effective manner; on the other hand, the application of SAR (Synthetic Aperture Radar) data in crop lodging monitoring is mostly oriented to the plot scale and has limitations in pixel scale or larger-scale applications. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, system, electronic device and medium for remote sensing monitoring of corn lodging, which can monitor the lodging situation of corn in a large area and effectively improve the monitoring accuracy.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] On the one hand, the present invention provides a method for remote sensing monitoring of corn lodging, including:

[0006] Obtaining field sample points of corn with different lodging degrees through field sampling and extracting Sentinel data at the modeling sample points; the corn with different lodging degrees includes normal non-lodged corn, slightly lodged corn and severely lodged corn; the Sentinel data includes Sentinel-1 radar data and Sentinel-2 optical image data;

[0007] Screening and preprocessing the Sentinel data at the modeling sample points to obtain preprocessed Sentinel data; the preprocessed Sentinel data includes full-time Sentinel-1 dual-polarization data and Sentinel-2 optical image data before lodging;

[0008] Calculate the normalized difference vegetation index (NDVI) of maize before lodging based on the Sentinel-2 optical image data before lodging;

[0009] Calculate the difference in the optimal sensitive parameters of maize during the disaster period and the recovery period based on the full-time Sentinel-1 dual-polarization data;

[0010] Divide all the modeling sample points into three levels: low vitality, medium vitality, and high vitality according to the NDVI of maize before lodging, and respectively calculate the mean value of the difference in the optimal sensitive parameters of maize during the disaster period and the recovery period at each level as the statistical threshold;

[0011] Establish a maize lodging monitoring decision tree based on the statistical threshold;

[0012] Determine the maize lodging degree in the monitored maize planting area according to the maize lodging monitoring decision tree.

[0013] Optionally, the field sample points with maize of different lodging degrees are obtained through field sampling, and the Sentinel data at the modeling sample points are extracted, specifically including:

[0014] Conduct field sampling on the plots where the normal non-lodged maize, slightly lodged maize, and severely lodged maize are located respectively as the field sample points;

[0015] Use the crop type distribution data of the study area to mask the Sentinel data of the field sample points, and extract the Sentinel data of all maize planting areas in the study area as the Sentinel data at the modeling sample points.

[0016] Optionally, the screening and preprocessing of the Sentinel data at the modeling sample points are carried out to obtain the preprocessed Sentinel data, specifically including:

[0017] Select the red band and near-infrared band data of the Sentinel-2 optical image data before lodging at the modeling sample points, and conduct cloud amount screening to obtain the Sentinel-2 optical image data before lodging;

[0018] Conduct time series analysis of the backscattering coefficients of the three polarization modes of VV, VH, and VH / VV of the Sentinel-1 radar data at the modeling sample points within the full-time phase period, select the VH polarization band data of the radar data, and conduct thermal noise removal, radiometric calibration, terrain correction, and filtering processing with a 5*5 window to obtain the full-time Sentinel-1 dual-polarization data.

[0019] Optionally, the calculation of the normalized difference vegetation index (NDVI) of maize before lodging based on the Sentinel-2 optical image data before lodging specifically includes:

[0020] Based on the Sentinel-2 optical image data before lodging, the normalized difference vegetation index NDVI of corn before lodging is calculated using the formula NDVI = (ρ NIR -ρ R ) / (ρ NIR +ρ R ); where ρ NIR and ρ g are the near-infrared band and the red band in the Sentinel-2 optical image data before lodging, respectively.

[0021] Optionally, calculating the difference in the best sensitive parameters during the disaster period and the recovery period of corn based on the all-time Sentinel-1 dual-polarization data specifically includes:

[0022] Based on the all-time Sentinel-1 dual-polarization data, the formula Δσ VH =σ VH_倒伏期 -σVH_before lodging is used to calculate the difference in the best sensitive parameters Δσ VH during the disaster period of corn; σ VH_倒伏期 represents the backscattering coefficient of the VH polarization band during the lodging period in the all-time Sentinel-1 dual-polarization data; σ VH_倒伏前 represents the backscattering coefficient of the VH polarization band before lodging in the all-time Sentinel-1 dual-polarization data;

[0023] Based on the all-time Sentinel-1 dual-polarization data, the formula δσ VH =σ VH_恢复期 -σ VH_倒伏期 is used to calculate the difference in the best sensitive parameters δσ VH during the recovery period of corn; where σ VH_恢复期 represents the backscattering coefficient of the VH polarization band during the recovery period in the all-time Sentinel-1 dual-polarization data.

[0024] Optionally, dividing all the modeling sample points into three levels of low vitality, medium vitality, and high vitality according to the normalized difference vegetation index of corn before lodging specifically includes:

[0025] If the normalized difference vegetation index NDVI of the corn at the modeling sample point before lodging < 0.4, the modeling sample point is divided into the low vitality level; if 0.4 < NDVI < 0.6, the modeling sample point is divided into the medium vitality level; if NDVI > 0.6, the modeling sample point is divided into the high vitality level.

[0026] On the other hand, the present invention also provides a remote sensing monitoring system for corn lodging, including:

[0027] The modeling sample point sentinel data extraction module is used to obtain field sample points of corn with different lodging degrees through field sampling and extract the sentinel data at the modeling sample points; the corn with different lodging degrees includes normal non-lodged corn, slightly lodged corn, and severely lodged corn; the sentinel data includes Sentinel-1 radar data and Sentinel-2 optical image data;

[0028] The screening and preprocessing module is used to screen and preprocess the sentinel data at the modeling sample points to obtain the preprocessed sentinel data; the preprocessed sentinel data includes full-time Sentinel-1 dual-polarization data and pre-lodging Sentinel-2 optical image data;

[0029] The normalized difference vegetation index calculation module is used to calculate the normalized difference vegetation index of corn before lodging according to the pre-lodging Sentinel-2 optical image data;

[0030] The optimal sensitive parameter difference calculation module is used to calculate the difference between the optimal sensitive parameters of corn during the disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data;

[0031] The threshold statistical analysis module is used to divide all modeling sample points into three levels: low vitality, medium vitality, and high vitality according to the normalized difference vegetation index of corn before lodging, and respectively statistically calculate the mean value of the difference between the optimal sensitive parameters of corn during the disaster period and the recovery period at each level as the statistical threshold;

[0032] The corn lodging monitoring decision tree establishment module is used to establish a corn lodging monitoring decision tree according to the statistical threshold;

[0033] The corn lodging degree division module is used to determine the corn lodging degree of the monitored corn planting area according to the corn lodging monitoring decision tree.

[0034] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the corn lodging remote sensing monitoring method described above is implemented.

[0035] On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed, the corn lodging remote sensing monitoring method described above is implemented.

[0036] According to the specific embodiments provided by the present invention, the following technical effects are disclosed by the present invention:

[0037] The present invention provides a method, system, electronic device and medium for remote sensing monitoring of maize lodging. The method includes: obtaining field sample points of maize with different lodging degrees through field sampling and extracting Sentinel data at the modeling sample points; screening and preprocessing the Sentinel data at the modeling sample points to obtain preprocessed Sentinel data; the preprocessed Sentinel data includes full-time Sentinel-1 dual-polarization data and Sentinel-2 optical image data before lodging; calculating the normalized difference vegetation index of maize before lodging according to the Sentinel-2 optical image data before lodging; calculating the difference in the best sensitive parameters of maize during the disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data; dividing all modeling sample points into three levels of low vigor, medium vigor and high vigor according to the normalized difference vegetation index of maize before lodging, and respectively counting the mean value of the difference in the best sensitive parameters of maize during the disaster period and the recovery period under each level as a statistical threshold; establishing a maize lodging monitoring decision tree according to the statistical threshold; and determining the maize lodging degree of the monitored maize planting area according to the maize lodging monitoring decision tree. The present invention uses the Sentinel-2 optical image data before lodging to monitor the crop plant vigor before maize lodging, and combines the full-time Sentinel-1 radar data to analyze the SAR polarization characteristics, so as to monitor the maize lodging situation in a large area, effectively improving the monitoring accuracy and providing a reliable reference for improving maize yield and maize lodging disaster warning. Description of the Drawings

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0039] Figure 1 It is a flowchart of the method for remote sensing monitoring of maize lodging provided by the embodiment of the present invention;

[0040] Figure 2 It is a technical roadmap of the method for remote sensing monitoring of maize lodging provided by the embodiment of the present invention;

[0041] Figure 3 It is a schematic diagram of the maize lodging monitoring decision tree established by the present invention;

[0042] Figure 4 It is a schematic diagram of the confusion matrix between the lodging degree classification result generated by the method of the present invention and the field verification points;

[0043] Figure 5 It is a partial enlarged view of the lodging degree classification result generated by the method of the present invention. Detailed Embodiments

[0044] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0045] The object of the present invention is to provide a method, system, electronic device and medium for remote sensing monitoring of corn lodging, which can monitor the lodging situation of corn in a large area and effectively improve the monitoring accuracy.

[0046] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0047] Figure 1 It is a flowchart of the method for remote sensing monitoring of corn lodging provided by the embodiments of the present invention. See Figure 1 The method for remote sensing monitoring of corn lodging in the present invention includes:

[0048] Step 101: Obtain field sample points of corn with different lodging degrees through field sampling and extract Sentinel data at the modeling sample points.

[0049] The present invention mainly monitors the lodging situation of corn after a typhoon passes through. The main influence period focuses on the jointing stage, tasseling stage and maturity stage of corn. The jointing stage refers to the period when the internodes of the stem of corn grow rapidly upward during the growth process. The tasseling stage refers to the stage when the male spike is extracted during the heading stage of corn. The maturity stage refers to the period when the dry matter of corn kernels is formed and rapidly accumulates. Corn with different lodging degrees includes normal non-lodged corn, slightly lodged corn and severely lodged corn.

[0050] The technical route of the method for remote sensing monitoring of corn lodging in the present invention is as Figure 2 shown. First, conduct field sampling on the corn fields in the wild, and conduct field sampling on the fields where normal non-lodged corn, slightly lodged corn and severely lodged corn are located respectively as field sample points. Then, use the crop type distribution data in the study area to mask the Sentinel data of the field sample points, extract the Sentinel data of the corn planting area in the study area, and use the extracted corn planting area in the study area as the modeling sample points.

[0051] The present invention uses Sentinel-1 radar data and Sentinel-2 optical images at the modeling sample points as the remote sensing data sources to obtain the remote sensing data during the maize lodging period in real time. The present invention combines optical images and radar data to study maize lodging. Through long-term data feature analysis, the SAR image information of maize from the tasseling stage to the maturity stage is analyzed, and the maize lodging period is divided into three periods: the pre-lodging period - the lodging occurrence period - the lodging recovery period. Among them, the pre-lodging period (abbreviated as pre-lodging) refers to the normal growth state period of maize before encountering bad weather such as typhoons. The lodging occurrence period (abbreviated as lodging period) refers to the period from when maize encounters extreme weather to the end of the extreme weather. The lodging recovery period (abbreviated as recovery period) refers to the stage when maize resumes growth after the extreme weather ends. Then the full time phase refers to the complete coverage of all time phase cycles of the pre-lodging period, the lodging occurrence period, and the lodging recovery period of maize.

[0052] Step 102: Screen and preprocess the Sentinel data at the modeling sample points to obtain the preprocessed Sentinel data; the preprocessed Sentinel data includes the full-time phase Sentinel-1 dual-polarization data and the pre-lodging Sentinel-2 optical image data.

[0053] The present invention uses Sentinel-1 radar data and Sentinel-2 optical images at the modeling sample points as the remote sensing data sources to obtain the remote sensing data during the maize lodging period in real time. Among them, the Sentinel-1 radar data has 2 polarization modes, including VV and VH, and the spatial resolution is 10 meters; the Sentinel-2 optical image data has 13 bands, among which the blue, green, red, and near-infrared four bands have a relatively high spatial resolution of 10 meters. The polarization modes of the Sentinel-1 radar data are divided into H horizontal polarization and V vertical polarization; VV means for vertical transmission and vertical reception, and VH means for vertical transmission and horizontal reception.

[0054] During the crop lodging monitoring process, the present invention selects the red and near-infrared bands of the optical image data and the VH polarization band of the radar data.

[0055] And to solve the problem of too many noise spots in the backscattering coefficient of the Sentinel-1 radar data, the Sentinel-1 radar data is subjected to thermal noise removal, radiometric calibration, terrain correction, and filtering processing with a 5*5 window to minimize the influence of the original data error on the results. The cloud amount of the Sentinel-2 optical image data is screened to effectively avoid the influence of too much cloud amount.

[0056] Therefore, the step 102 screens and preprocesses the Sentinel data at the modeling sample points to obtain the preprocessed Sentinel data, which specifically includes:

[0057] Select the red band and near-infrared band data of the Sentinel-2 optical image data before lodging at the modeling sample points, and perform cloud cover screening to obtain the Sentinel-2 optical image data before lodging.

[0058] Perform time series analysis of the backscattering coefficients of the three polarization modes of VV, VH, and VH / VV in the Sentinel-1 radar data at the modeling sample points within the full-time phase period. Select the VH polarization band data of the radar data and perform thermal noise removal, radiometric calibration, terrain correction, and filtering processing with a 5*5 window to obtain the full-time phase Sentinel-1 dual-polarization data.

[0059] Step 103: Calculate the normalized difference vegetation index of maize before lodging based on the Sentinel-2 optical image data before lodging.

[0060] Calculate the NDVI of the maize planting area from the Sentinel-2 optical data according to formula (1), and extract the normalized difference vegetation index of maize before lodging at the modeling sample points.

[0061] The calculation formula of the Normalized Difference Vegetation Index (NDVI) is as follows:

[0062] NDVI = (ρ NIR - ρ R ) / (ρ NIR + ρ R ) (1)

[0063] Where NDVI is the normalized difference vegetation index of maize before lodging, and ρ NIR and ρ R are the near-infrared band and red band in the Sentinel-2 optical image data before lodging, respectively.

[0064] Step 104: Calculate the difference in the best sensitive parameters of maize during the disaster period and the recovery period based on the full-time phase Sentinel-1 dual-polarization data.

[0065] Based on Google Earth Engine (GEE), perform time series analysis of the backscattering coefficients of the three polarization modes of VV, VH, and VH / VV in the Sentinel-1 radar data at the modeling sample points within the full-time phase period. From the change rules of the backscattering coefficients before and after lodging of samples with different lodging degrees, it can be seen that with the occurrence of lodging, the backscattering coefficients all show an increasing trend. Among them, the VH polarization is highly sensitive to lodging and has good separability for different lodging degrees. Therefore, the present invention selects the backscattering coefficient of VH polarization as the best sensitive parameter (OSP) for subsequent analysis and calculation of threshold parameters. That is to say, the data required for modeling in the present invention are the full-time phase Sentinel-1 dual-polarization data and the Sentinel-2 optical image data before lodging.

[0066] Take the pre - lodging NDVI of corn calculated from the Sentinel - 2 optical image data before lodging by the above formula (1) as an indicator factor of the pre - lodging crop plant vigor. By studying and setting the empirical threshold of NDVI, NDVI < 0.4 is low vigor, 0.4 < NDVI < 0.6 is medium vigor, and NDVI > 0.6 is high vigor. Thus, the corn planting area is divided into three vigor levels: low, medium, and high. Calculate the difference of the VH backscattering coefficient of the Sentinel - 1 dual - polarization data at three periods: before lodging, during lodging, and during recovery. The calculation methods are shown in formulas (2) and (3). Then, use the decision - tree analysis method to establish a corn lodging monitoring model, so as to realize the division of the corn lodging degree.

[0067] Δσ OSP = Δσ VH = σ VH_倒伏期 - σ VH_倒伏前 (2)

[0068] δσ OSP = δσ VH = σ VH_烣复期 - σ VH_倒伏期 (3)

[0069] Where the difference Δσ VH and the difference δσ VH during the disaster period and the recovery period of corn are the differences of the best - sensitive parameters during the disaster period and the recovery period of corn respectively, and can also be denoted as Δσ OSP and δσ OSP . σ VH represents the backscattering coefficient of the VH polarization band; specifically, σ VH_倒伏前 , σ VH_倒伏期 , σ VH_恢复期 represent the backscattering coefficients of the VH polarization band before lodging, during lodging, and during recovery in the Sentinel - 1 dual - polarization data of the full time - phase respectively.

[0070] Step 105: Divide all the modeling sample points into three levels: low vigor, medium vigor, and high vigor according to the pre - lodging normalized vegetation index of the corn, and statistically calculate the mean value of the differences of the best - sensitive parameters during the disaster period and the recovery period of the corn under each level as the statistical threshold.

[0071] The method for dividing the NDVI crop vigor level of the modeling sample points is as follows: If the pre - lodging normalized vegetation index NDVI of the modeling sample point is < 0.4, the modeling sample point is divided into the low - vigor level; if 0.4 < NDVI < 0.6, the modeling sample point is divided into the medium - vigor level; if NDVI > 0.6, the modeling sample point is divided into the high - vigor level.

[0072] By analyzing the Δσ VHand δσ VH It was found that under different vigor levels, as the lodging degree increased, Δσ VH and δσ VH changed differently. Thus, a maize lodging monitoring decision tree as shown in Figure 3 could be established as a maize lodging monitoring model. First, all maize planting areas of the modeling sample points were divided into three levels: low vigor, medium vigor, and high vigor using the NDVI value. At each level, the mean (avg) and covariance (std) of Δσ VH and δσ VH were statistically analyzed. For the convenience of decision tree representation, the means of Δσ VH and δσ VH were respectively denoted as D and R. The statistical results are shown in Table 1.

[0073] Table 1 Statistical Thresholds for Maize Lodging Degree Classification

[0074] Vigor of maize plants <![CDATA[avg-Δσ VH (D)]]> <![CDATA[std-Δσ VH > <![CDATA[avg-δσ VH (R)]]> <![CDATA[std-δσ VH > Low vigor 0.086 0.921 -0.050 1.237 Medium vigor 0.269 0.674 -0.171 1.013 High vigor 0.574 0.614 -0.821 0.901

[0075] The D and R values corresponding to low, medium, and high vigor in Table 1 are d1, d2, d3 and r1, r2, r3 in the maize lodging monitoring decision tree, that is, the statistical thresholds required for establishing the maize lodging monitoring decision tree of the present invention. Based on these statistical thresholds, maize can be classified into normal, slightly lodged, and severely lodged. Figure 3

[0076] Step 106: Establish a maize lodging monitoring decision tree according to the statistical thresholds.

[0077] Take the NDVI of maize before lodging calculated from the Sentinel-2 optical image data before lodging by formula (1) as the indicator factor of crop plant vigor before lodging, divide all modeling sample points into three levels: low vigor, medium vigor, and high vigor, and take the D and R values corresponding to low, medium, and high vigor as the statistical thresholds d1, d2, d3 and r1, r2, r3 in the maize lodging monitoring decision tree, and establish a maize lodging monitoring decision tree as shown in Figure 3

[0078] Step 107: Determine the maize lodging degree of the monitored maize planting area according to the maize lodging monitoring decision tree.

[0079] When classifying the maize lodging degree of the monitored maize planting area, it is necessary to first calculate the NDVI of maize before lodging in the monitored maize planting area according to formula (1), and calculate the D and R values in the monitored maize planting area according to formulas (2) and (3), and then substitute the NDVI of maize before lodging and the D and R values in the monitored maize planting area into Figure 3 ​​In the shown maize lodging monitoring decision tree, by comparing with the pre-trained NDVI statistical thresholds (0.4, 0.6) and D and R statistical thresholds (d1, d2, d3, r1, r2, r3), according to the comparison logic in the maize lodging monitoring decision tree, the maize in the monitored maize planting area can be classified into normal, slightly lodged, and severely lodged, serving as the classification result of the maize lodging degree in the monitored maize planting area.

[0080] For example, referring to Figure 3 , when the NDVI in the monitored maize planting area > 0.6, if D > d3 and R > r3, it can be determined that the maize in the monitored maize planting area is severely lodged; if D > d3 and R ≤ r3, it can be determined that the maize in the monitored maize planting area is slightly lodged; if D ≤ d3, it can be determined that the maize in the monitored maize planting area is in a normal state. When 0.4 < NDVI ≤ 0.6 in the monitored maize planting area, if D ≤ d2, it can be determined that the maize in the monitored maize planting area is in a normal state; if D > d2 and R ≤ r2, it can be determined that the maize in the monitored maize planting area is slightly lodged; if D > d2 and R > r2, it can be determined that the maize in the monitored maize planting area is severely lodged. When the NDVI in the monitored maize planting area ≤ 0.4, if D ≤ d1, it can be determined that the maize in the monitored maize planting area is in a normal state; if D > d1 and R ≤ r1, it can be determined that the maize in the monitored maize planting area is slightly lodged; if D > d1 and R > r1, it can be determined that the maize in the monitored maize planting area is severely lodged.

[0081] The present invention combines optical images and radar data to study maize lodging. Through long-term data feature analysis, the SAR image information of maize from the tasseling stage to the maturity stage is analyzed. The maize lodging cycle is divided into three periods: pre-lodging period - lodging occurrence period - lodging recovery period. By using the optical image data before lodging to monitor the crop plant vigor before maize lodging and combining with the SAR images in each cycle of the full time phase, the lodging situation of maize in a large area can be monitored, effectively improving the monitoring accuracy and providing a reliable reference for increasing maize yield and early warning of maize lodging disasters.

[0082] Next, a specific embodiment is used to verify the accuracy of the maize lodging remote sensing monitoring method of the present invention.

[0083] The maize lodging remote sensing monitoring method proposed by the present invention is tested in Northeast China (14 prefecture-level cities in Jilin Province and Heilongjiang Province). Figure 4 The confusion matrix of the lodging degree classification result of the method of the present invention and the field verification points, referring to Figure 4, in terms of accuracy verification, the accuracy of normal corn is 91.7%, the accuracy of mild lodging is 93.3%, and the overall accuracy of severe lodging can reach 94.2%, indicating that the method of the present invention has high monitoring accuracy. Figure 5 It is a partial enlarged view of the lodging monitoring result generated by the method of the present invention. See Figure 5 , from a visual perspective, the classification result of the corn lodging degree divided by the method of the present invention has a high degree of coincidence with the plot information of the actual plot, indicating that the method of the present invention can monitor the lodging degree on a large regional scale.

[0084] Based on the method provided by the present invention, the present invention also provides a corn lodging remote sensing monitoring system, including:

[0085] A modeling sample point sentinel data extraction module for obtaining field sample points of corn with different lodging degrees through field sampling and extracting sentinel data at the modeling sample points; the corn with different lodging degrees includes normal non-lodged corn, mildly lodged corn, and severely lodged corn; the sentinel data includes Sentinel-1 radar data and Sentinel-2 optical image data;

[0086] A screening and preprocessing module for screening and preprocessing the sentinel data at the modeling sample points to obtain preprocessed sentinel data; the preprocessed sentinel data includes full-time Sentinel-1 dual-polarization data and pre-lodging Sentinel-2 optical image data;

[0087] A normalized difference vegetation index calculation module for calculating the pre-lodging normalized difference vegetation index of corn according to the pre-lodging Sentinel-2 optical image data;

[0088] An optimal sensitive parameter difference calculation module for calculating the difference between the optimal sensitive parameters during the disaster period and the recovery period of corn according to the full-time Sentinel-1 dual-polarization data;

[0089] A threshold statistical analysis module for dividing all modeling sample points into three levels of low vitality, medium vitality, and high vitality according to the pre-lodging normalized difference vegetation index of corn, and respectively statistically calculating the mean value of the difference between the optimal sensitive parameters during the disaster period and the recovery period of corn at each level as the statistical threshold;

[0090] A corn lodging monitoring decision tree establishment module for establishing a corn lodging monitoring decision tree according to the statistical threshold;

[0091] A corn lodging degree division module for determining the corn lodging degree of the monitored corn planting area according to the corn lodging monitoring decision tree.

[0092] The present invention is directed to corn after the jointing stage. By using multi-temporal Sentinel data, the SAR polarization characteristics of a long time series before and after corn lodging are analyzed. A monitoring method for corn lodging in extreme weather is established by combining pre-disaster crop vitality and SAR polarization characteristics. Since the remote sensing monitoring method for corn lodging in the present invention uses long-term Sentinel data, it can not only meet the characteristic changes in the time period but also ensure the spatial resolution of the data. Moreover, the method of the present invention combines optical images with radar data. On the premise of considering the plant vitality of corn before the disaster, the SAR polarization characteristics are further analyzed, thereby establishing a corn lodging monitoring decision tree. According to this corn lodging monitoring decision tree, corn can be classified into normal, slightly lodged, and severely lodged, which greatly improves the monitoring accuracy and can monitor the lodging degree on a large regional scale, having broad application prospects.

[0093] Furthermore, the present invention also provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The processor can call the computer program in the memory to execute the remote sensing monitoring method for corn lodging.

[0094] In addition, when the computer program in the above-mentioned memory is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories, random access memories, magnetic disks, or optical discs that can store program codes.

[0095] Furthermore, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed, it can implement the remote sensing monitoring method for corn lodging.

[0096] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0097] In this article, specific examples are used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A remote sensing monitoring method for maize lodging, characterized in that, Including: Obtaining field sample points of maize with different lodging degrees through field sampling and extracting sentinel data at the modeling sample points; the maize with different lodging degrees includes normal non-lodged maize, slightly lodged maize, and severely lodged maize; the sentinel data includes Sentinel-1 radar data and Sentinel-2 optical image data. Screening and preprocessing the sentinel data at the modeling sample points to obtain preprocessed sentinel data; the preprocessed sentinel data includes full-time Sentinel-1 dual-polarization data and pre-lodging Sentinel-2 optical image data. Calculating the pre-lodging normalized difference vegetation index of maize according to the pre-lodging Sentinel-2 optical image data. Calculating the difference in the best sensitive parameters of maize during the disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data. The calculating the difference in the best sensitive parameters of maize during the disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data specifically includes: According to the Sentinel-1 full-time dual-polarization data, the formula Δσ VH = σ VH_倒伏期 - σ VH_倒伏前 is used to calculate the difference Δσ VH of the optimal sensitivity parameters during the maize disaster period; σ VH_倒伏期 represents the backscattering coefficient of the VH polarization band during the lodging period in the Sentinel-1 full-time dual-polarization data; σ VH_倒伏前 represents the backscattering coefficient of the VH polarization band before lodging in the Sentinel-1 full-time dual-polarization data; According to the Sentinel-1 dual-polarization data of the full temporal phase, the formula δσ VH = σ VH_恢复期 - σ VH_倒伏期 is used to calculate the difference δσ VH of the optimal sensitive parameters during the corn recovery period; where δ VH_恢复期 represents the backscattering coefficient of the VH polarization band during the recovery period in the Sentinel-1 dual-polarization data of the full temporal phase; Dividing all modeling sample points into three levels of low vitality, medium vitality, and high vitality according to the pre-lodging normalized difference vegetation index of maize, and respectively counting the mean values of the differences in the best sensitive parameters of maize during the disaster period and the recovery period under each level as statistical thresholds. Establishing a maize lodging monitoring decision tree according to the statistical thresholds. Determining the lodging degree of maize in the monitored maize planting area according to the maize lodging monitoring decision tree.

2. The maize lodging remote sensing monitoring method according to claim 1, wherein The obtaining field sample points of maize with different lodging degrees through field sampling and extracting sentinel data at the modeling sample points specifically includes: Conducting field sampling on the plots where normal non-lodged maize, slightly lodged maize, and severely lodged maize are located respectively as field sample points. Using the crop type distribution data of the study area to mask the sentinel data of the field sample points, and extracting the sentinel data of all maize planting areas in the study area as the sentinel data at the modeling sample points.

3. The maize lodging remote sensing monitoring method according to claim 1, characterized in that The screening and preprocessing the sentinel data at the modeling sample points to obtain preprocessed sentinel data specifically includes: Selecting the red band and near-infrared band data of the pre-lodging Sentinel-2 optical image data at the modeling sample points and conducting cloud amount screening to obtain the pre-lodging Sentinel-2 optical image data. Performing time series analysis on the backscattering coefficients of the VV, VH, and VH / VV polarization modes in the Sentinel-1 radar data at the modeling sample points within the full-time phase period, selecting the VH polarization band data of the radar data and conducting thermal noise removal, radiometric calibration, terrain correction, and filtering processing with a 5*5 window to obtain full-time Sentinel-1 dual-polarization data.

4. The maize lodging remote sensing monitoring method according to claim 1, characterized in that, The calculating the pre-lodging normalized difference vegetation index of maize according to the pre-lodging Sentinel-2 optical image data specifically includes: According to the Sentinel-2 optical image data before lodging, the normalized difference vegetation index NDVI of maize before lodging is calculated using the formula NDVI = (ρ NIR - ρ R ) / (ρ NIR + ρ R ); where ρ NIR and ρ R are the near-infrared band and the red band in the Sentinel-2 optical image data before lodging, respectively.

5. The maize lodging remote sensing monitoring method according to claim 4, wherein, The dividing all modeling sample points into three levels of low vitality, medium vitality, and high vitality according to the pre-lodging normalized difference vegetation index of maize specifically includes: If the normalized difference vegetation index (NDVI) of the modeling sample points before corn lodging is less than 0.4, the modeling sample points are classified into the low-vitality level; if 0.4 < NDVI < 0.6, the modeling sample points are classified into the medium-vitality level; if NDVI > 0.6, the modeling sample points are classified into the high-vitality level.

6. A remote sensing monitoring system for corn lodging, characterized in that, Including: A sentinel data extraction module for modeling sample points, which is used to obtain field sample points of corn with different lodging degrees through field sampling and extract the sentinel data at the modeling sample points; the corn with different lodging degrees includes normal non-lodged corn, slightly lodged corn, and severely lodged corn; the sentinel data includes Sentinel-1 radar data and Sentinel-2 optical image data; A screening and preprocessing module, which is used to screen and preprocess the sentinel data at the modeling sample points to obtain the preprocessed sentinel data; the preprocessed sentinel data includes full-time Sentinel-1 dual-polarization data and Sentinel-2 optical image data before lodging. A normalized difference vegetation index calculation module, which is used to calculate the normalized difference vegetation index of corn before lodging according to the Sentinel-2 optical image data before lodging. An optimal sensitive parameter difference calculation module, which is used to calculate the difference between the optimal sensitive parameters of the corn disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data. The calculation of the difference between the optimal sensitive parameters of the corn disaster period and the recovery period according to the full-time Sentinel-1 dual-polarization data specifically includes: Based on the full - time Sentinel - 1 dual - polarization data, the formula Δσ VH = σ VH_倒伏期 - σ VH_倒伏前 is used to calculate the difference in the best sensitive parameters Δσ VH during the maize disaster period; σ VH_倒伏期 represents the backscattering coefficient of the VH polarization band during the lodging period in the full - time Sentinel - 1 dual - polarization data; σ VH_倒伏前 represents the backscattering coefficient of the VH polarization band before lodging in the full - time Sentinel - 1 dual - polarization data; According to the full-time Sentinel-1 dual-polarization data, the formula δσ VH = σ VH_恢复期 - σ VH_倒伏期 is used to calculate the difference δσ VH of the optimal sensitive parameters during the corn recovery period; where σ VH_恢复期 represents the backscattering coefficient of the VH polarization band during the recovery period in the full-time Sentinel-1 dual-polarization data; A threshold statistical analysis module, which is used to classify all modeling sample points into three levels: low vitality, medium vitality, and high vitality according to the normalized difference vegetation index of corn before lodging, and respectively calculate the mean value of the difference between the optimal sensitive parameters of the corn disaster period and the recovery period under each level as the statistical threshold. A corn lodging monitoring decision tree establishment module, which is used to establish a corn lodging monitoring decision tree according to the statistical threshold. A corn lodging degree classification module, which is used to determine the corn lodging degree of the monitored corn planting area according to the corn lodging monitoring decision tree.

7. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the corn lodging remote sensing monitoring method according to any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the corn lodging remote sensing monitoring method according to any one of claims 1 to 5.

Images