A method for rice carbon sink inversion by fusing unmanned aerial vehicle photogrammetry and sentinel 1 satellite vv polarization coordination

Abstract

This invention discloses a method for rice carbon sequestration retrieval that integrates UAV photogrammetry and Sentinel-1 satellite VV polarization synergy, belonging to the field of agricultural remote sensing monitoring technology. The method includes image acquisition, image data processing, image pixel spatial alignment, rice sampling, rice sample processing, sample data processing, carbon sequestration and fresh weight fitting, carbon sequestration per unit area and water content per unit area fitting at sampling points, modeling, feature image retrieval, comparison and screening, paddy field soil elevation determination, rice height distribution thematic map creation, height carbon sequestration fitting, height-based carbon sequestration retrieval, and synergistic carbon sequestration retrieval, ultimately obtaining a rice carbon sequestration thematic map. This invention achieves rice carbon sequestration retrieval based on UAV photogrammetry and Sentinel-1 satellite VV polarization synergy, improving the monitoring accuracy of carbon sequestration in rice plants, especially rice straw.

Description

Technical Field

[0001] This invention relates to the field of agricultural remote sensing monitoring technology, and in particular to a method for inverting rice carbon sequestration by integrating UAV photogrammetry and Sentinel-1 satellite VV polarization coordination. Background Technology

[0002] Currently, mainstream carbon sequestration monitoring mainly focuses on the forestry sector, and the monitoring technology is primarily optical, mainly based on volumetric measurement methods such as lidar for single monitoring. Agricultural carbon sequestration is receiving increasing attention, but currently, it is largely concentrated on soil organic matter carbon sequestration. Rice roots and straw are also important components of agricultural carbon sequestration; at the same time, the burning of straw is a major source of carbon emissions. Overall, research on carbon sequestration monitoring of rice plants is limited. Summary of the Invention

[0003] In view of this, the present invention provides a method for rice carbon sequestration inversion that integrates UAV photogrammetry and Sentinel-1 satellite VV polarization coordination, realizing a rapid carbon sequestration inversion method for rice just before harvest. To achieve the above objectives, the present invention adopts the following technical solution: A method for retrieving rice carbon sequestration by integrating UAV photogrammetry and Sentinel-1 satellite VV polarization synergy, characterized by the following steps: S1 image acquisition steps: Select the sampling area and acquire the UAV digital elevation model of the paddy field and Sentinel-1 radar satellite VV polarization image on the same day before the rice harvest. S2 image data processing steps: Perform geometric correction, radiometric calibration and other preprocessing on the VV polarization images of Sentinel-1 radar satellite to obtain backscattering coefficient feature images. Based on bicubic interpolation, improve the resolution of the backscattering coefficient feature images to 1 meter. Obtain the UAV rice paddy digital elevation model and its plane coordinates. Take the average value of all elevation data within 1 meter range as the elevation value H of the pixel and reconstruct the 1-meter resolution digital elevation model. S3 Image Pixel Spatial Alignment Steps: The digital elevation model of the same resolution and the backscattering coefficient feature image are registered and aligned in planar spatial coordinates to obtain TIFF image data with unified coordinates and pixel registration and alignment. S4 Rice Sampling Procedure: On the same day as image acquisition, at least 50 sampling points are randomly set up and their coordinates are marked using an RTK device. Within a 1-meter diameter range of each sampling point, the row spacing a, pier spacing d, and the number of rice plants n per pier are randomly selected for measurement. Three rice plants are selected to measure their height and cut along the soil surface. The samples are sealed in plastic bags and labeled with sample numbers. S5 rice sample processing steps: Weigh the fresh weight of each cut rice sample, dry it and weigh the dry weight, and finally use a carbonization furnace to perform anaerobic carbonization and weigh the remaining carbonized components as the rice carbon sequestration. S6 sampling data processing steps: Take the average parameters of 3 rice samples at the same sampling point to obtain the average rice height h, average fresh weight W, average dry weight G, average carbon sink C, average carbon content γ of average fresh weight, average water content WC, number of rice plants per unit area N, water content per unit area VWC, and carbon sink VC of the sampling point. S7 Carbon sink and fresh weight fitting steps: Perform statistical analysis on the average fresh weight of all sample points and the rice carbon sink data to obtain a linear relationship: C = W×γ, and calculate the coefficient of determination DC1. The steps for fitting carbon sequestration per unit area and water content per unit area at sampling point S8 are as follows: Statistical analysis is performed on carbon sequestration per unit area (VC) and water content per unit area (VWC) to obtain the fitting relationship: VC = VWC × k + l, with a coefficient of determination DC2, where k and l are constant parameters. S9 modeling steps: Randomly select half of the sampling points and extract the backscattering coefficient feature image pixel value sigma0 based on the coordinates. Quantitatively model the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and establish a quantitative model sigma0=F(UW) and sigma0=F(VWC). S10 Feature Image Inversion Steps: Substitute the backscattering coefficient feature image pixel values ​​corresponding to the remaining samples into the quantitative models sigma0=F(UW) and sigma0=F(VWC) to obtain the inversion values ​​of fresh weight per unit area UW' and water content per unit area VWC' of the remaining sampling points. Compare these values ​​with the measured values ​​to obtain the coefficient of determination DC3 between the inversion of fresh weight per unit area and the measured values, and the coefficient of determination DC4 between the measured and measured values ​​of water content per unit area. S11 Comparison and Screening Steps: Compare the sizes of DC1×DC3 and DC2×DC4. If DC1×DC3 > DC2×DC4, substitute each pixel value of the backscattering coefficient feature image into sigma0=F(UW) to calculate the fresh weight per unit area UW' of each pixel. Then, use the model VC=UW×γ to iterate through each pixel to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. Otherwise, substitute each pixel value of the backscattering coefficient feature image into the model sigma0=F(VWC) to calculate the water content per unit area VWC' of each pixel. Then, use the model VC=VWC×k+l to iterate through each pixel to obtain the fresh weight per unit area to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. S12 Paddy Field Soil Elevation Measurement Steps: Extract the elevation value H of each sampling point from the 1-meter resolution digital elevation model, and simultaneously extract the average rice height h of the sampling points. Calculate the soil elevation H0=Hh of each sampling point, and use linear interpolation to construct a 1-meter resolution soil digital elevation model registered with the paddy field digital elevation model. Steps for creating the S13 rice height distribution thematic map: Difference the pixel value of each cell in the registered 1-meter resolution paddy field digital elevation model and the soil digital elevation model to obtain the TIFF file of the paddy field rice height distribution thematic map. S14 High-level carbon sequestration fitting steps: Fit the rice height h corresponding to the sampling point with the carbon sequestration per unit area VC: VC=p×h+q, where p and q are constant terms. S15 Step for height inversion carbon sink: The fitted model VC=p×h+q is used to traverse each pixel of the rice height distribution thematic map TIFF file to obtain the height inversion carbon sink thematic map. S16 Co-inversion Carbon Sequestration Steps: After registering the 1-meter resolution radar-inverted rice carbon sequestration thematic map TIFF file with the height-inverted carbon sequestration thematic image pixels, the two carbon sequestration values ​​of each pixel are added together and averaged to obtain the co-inverted rice carbon sequestration thematic map TIFF file. Optionally, S1 in the above method includes the following steps: S1-1 Rice Paddy Drone Digital 3D Model Creation: Data of the experimental field was collected by drone aerial photography and a digital 3D model of the rice paddy was created using 2D Gaussian splashing technology. S1-2 Converts the created 3D digital model of the paddy field into a 2D digital elevation model TIFF file; Optionally, S6 of the above method includes the following steps: Steps for measuring the unit area moisture content at sampling point S6-1: Calculate the average plant moisture content WC by subtracting the average fresh weight from the average dry weight; calculate the number of rice plants per unit area at each sampling point: N=n / (a×d); finally, obtain the unit area moisture content at the sampling point: VWC=WC×N. S6-2 Sampling Point Carbon Sequestration Measurement Steps: Carbon Sequestration per Unit Area VC = C × N; Steps for calculating the fresh weight of rice per unit area at each sampling point in S6-3: Fresh weight of rice per unit area UW = W × N; Optionally, S9 of the above method includes the following steps: S9-1 Quantitative Fitting Steps: Extract the pixel value sigma0 of the backscattering coefficient feature image at 1-meter resolution and perform scatter analysis with the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and then perform least squares fitting. The advantages of this invention are: This technology uses two-dimensional Gaussian splashing technology and Sentinel-1 satellite VV polarization image data to improve the accuracy and efficiency of rice carbon sink inversion, and is stable and reliable with good application value. Attached Figure Description To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort. Figure 1 A schematic diagram of a method for retrieving rice carbon sequestration by integrating UAV photogrammetry and Sentinel-1 satellite VV polarization synergy; Detailed Implementation The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. See Figure 1 As shown, this invention discloses a method for inverting rice carbon sequestration by integrating UAV photogrammetry and Sentinel-1 satellite VV polarization synergy, comprising the following steps: S1 image acquisition steps: Select the sampling area and acquire the UAV digital elevation model of the paddy field and Sentinel-1 radar satellite VV polarization image on the same day before the rice harvest. S2 image data processing steps: Perform geometric correction, radiometric calibration and other preprocessing on the VV polarization images of Sentinel-1 radar satellite to obtain backscattering coefficient feature images. Based on bicubic interpolation, improve the resolution of the backscattering coefficient feature images to 1 meter. Obtain the UAV rice paddy digital elevation model and its plane coordinates. Take the average value of all elevation data within 1 meter range as the elevation value H of the pixel and reconstruct the 1-meter resolution digital elevation model. S3 Image Pixel Spatial Alignment Steps: The UAV photogrammetry and backscatter coefficient feature images of the same resolution are registered and aligned in planar spatial coordinates to obtain TIFF image data with unified coordinates and pixel registration and alignment. S4 Rice Sampling Procedure: On the same day as image acquisition, at least 50 sampling points are randomly set up and their coordinates are marked using an RTK device. Within a 1-meter diameter range of each sampling point, the row spacing a, pier spacing d, and the number of rice plants n per pier are randomly selected for measurement. Three rice plants are selected to measure their height and cut along the soil surface. The samples are sealed in plastic bags and labeled with sample numbers. S5 rice sample processing steps: Weigh the fresh weight of each cut rice sample, dry it and weigh the dry weight, and finally use a carbonization furnace to perform anaerobic carbonization and weigh the remaining carbonized components as the rice carbon sequestration. S6 sampling data processing steps: Take the average parameters of 3 rice samples at the same sampling point to obtain the average rice height h, average fresh weight W, average dry weight G, average carbon sink C, average carbon content γ of average fresh weight, average water content WC, number of rice plants per unit area N, water content per unit area VWC, and carbon sink VC of the sampling point. S7 Carbon sink and fresh weight fitting steps: Perform statistical analysis on the average fresh weight of all sample points and the rice carbon sink data to obtain a linear relationship: C = W×γ, and calculate the coefficient of determination DC1. The steps for fitting carbon sequestration per unit area and water content per unit area at sampling point S8 are as follows: Statistical analysis is performed on carbon sequestration per unit area (VC) and water content per unit area (VWC) to obtain the fitting relationship: VC = VWC × k + l, with a coefficient of determination DC2, where k and l are constant parameters. S9 modeling steps: Randomly select half of the sampling points and extract the backscattering coefficient feature image pixel value sigma0 based on the coordinates. Quantitatively model the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and establish a quantitative model sigma0=F(UW) and sigma0=F(VWC). S10 Feature Image Inversion Steps: Substitute the backscattering coefficient feature image pixel values ​​corresponding to the remaining samples into the quantitative models sigma0=F(UW) and sigma0=F(VWC) to obtain the inversion values ​​of fresh weight per unit area UW' and water content per unit area VWC' of the remaining sampling points. Compare these values ​​with the measured values ​​to obtain the coefficient of determination DC3 between the inversion of fresh weight per unit area and the measured values, and the coefficient of determination DC4 between the measured and measured values ​​of water content per unit area. S11 Comparison and Screening Steps: Compare the sizes of DC1×DC3 and DC2×DC4. If DC1×DC3 > DC2×DC4, substitute each pixel value of the backscattering coefficient feature image into sigma0=F(UW) to calculate the fresh weight per unit area UW' of each pixel. Then, use the model VC=UW×γ to traverse each pixel to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. Otherwise, substitute each pixel value of the backscattering coefficient feature image into the model sigma0=F(VWC) to calculate the water content per unit area VWC' of each pixel. Then, use the model VC=VWC×k+l to traverse each pixel to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. S12 Paddy Field Soil Elevation Measurement Steps: Extract the elevation value H of each sampling point from the 1-meter resolution digital elevation model, and simultaneously extract the average rice height h of the sampling points. Calculate the soil elevation H0=Hh of each sampling point, and use linear interpolation to construct a 1-meter resolution soil digital elevation model registered with the paddy field digital elevation model. Steps for creating the S13 rice height distribution thematic map: Difference the pixel value of each cell in the registered 1-meter resolution paddy field digital elevation model and the soil digital elevation model to obtain the TIFF file of the paddy field rice height distribution thematic map. S14 High-level carbon sequestration fitting steps: Fit the rice height h corresponding to the sampling point with the carbon sequestration per unit area VC: VC=p×h+q, where p and q are constant terms. S15 Step for height inversion carbon sink: The fitted model VC=p×h+q is used to traverse each pixel of the rice height distribution thematic map TIFF file to obtain the height inversion carbon sink thematic map. S16 Co-inversion Carbon Sequestration Steps: After registering the 1-meter resolution radar-inverted rice carbon sequestration thematic map TIFF file with the height-inverted carbon sequestration thematic image pixels, the two carbon sequestration values ​​of each pixel are added together and averaged to obtain the co-inverted rice carbon sequestration thematic map TIFF file. Optionally, S1 in the above method includes the following steps: S1-1 Rice Paddy Drone Digital 3D Model Creation: Data of the experimental field was collected by drone aerial photography and a digital 3D model of the rice paddy was created using 2D Gaussian splashing technology. S1-2 Converts the created 3D digital model of the paddy field into a 2D digital elevation model TIFF file; Optionally, S6 of the above method includes the following steps: Steps for measuring the unit area moisture content at sampling point S6-1: Calculate the average plant moisture content WC by subtracting the average fresh weight from the average dry weight; calculate the number of rice plants per unit area at each sampling point: N=n / (a×d); finally, obtain the unit area moisture content VWC=WC×N at the sampling point. S6-2 Sampling Point Carbon Sequestration Measurement Steps: Carbon Sequestration per Unit Area VC = C × N; Steps for calculating the fresh weight of rice per unit area at each sampling point in S6-3: Fresh weight of rice per unit area UW = W × N; Optionally, S9 of the above method includes the following steps: S9-1 Quantitative Fitting Steps: Extract the pixel value sigma0 of the backscattering coefficient feature image at 1-meter resolution and perform scatter analysis with the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and then perform least squares fitting. The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for inverting rice carbon sequestration by integrating UAV photogrammetry and Sentinel-1 satellite VV polarization synergy, characterized by the following steps: S1 image acquisition steps: Select the sampling area and acquire the UAV digital elevation model of the paddy field and Sentinel-1 radar satellite VV polarization image on the same day before the rice harvest. S2 image data processing steps: Perform geometric correction, radiometric calibration and other preprocessing on the VV polarization images of Sentinel-1 radar satellite to obtain backscattering coefficient feature images. Based on bicubic interpolation, improve the resolution of the backscattering coefficient feature images to 1 meter. Obtain the UAV rice paddy digital elevation model and its plane coordinates. Take the average value of all elevation data within 1 meter range as the elevation value H of the pixel and reconstruct the 1-meter resolution digital elevation model. S3 Image Pixel Spatial Alignment Steps: The digital elevation model of the same resolution and the backscattering coefficient feature image are registered and aligned in planar spatial coordinates to obtain TIFF image data with unified coordinates and pixel registration and alignment. S4 Rice Sampling Procedure: On the same day as image acquisition, at least 50 sampling points are randomly set up and their coordinates are marked using an RTK device. Within a 1-meter diameter range of each sampling point, the row spacing a, pier spacing d, and the number of rice plants n per pier are randomly selected for measurement. Three rice plants are selected to measure their height and cut along the soil surface. The samples are sealed in plastic bags and labeled with sample numbers. S5 rice sample processing steps: Weigh the fresh weight of each cut rice sample, dry it and weigh the dry weight, and finally use a carbonization furnace to perform anaerobic carbonization and weigh the remaining carbonized components as the rice carbon sequestration. S6 sampling data processing steps: Take the average parameters of 3 rice samples at the same sampling point to obtain the average rice height h, average fresh weight W, average dry weight G, average carbon sink C, average carbon content γ of average fresh weight, average water content WC, number of rice plants per unit area N, water content per unit area VWC, and carbon sink VC of the sampling point. S7 Carbon sink and fresh weight fitting steps: Perform statistical analysis on the average fresh weight of all sample points and the rice carbon sink data to obtain a linear relationship: C = W×γ, and calculate the coefficient of determination DC1. The steps for fitting carbon sequestration per unit area and water content per unit area at sampling point S8 are as follows: Statistical analysis is performed on carbon sequestration per unit area (VC) and water content per unit area (VWC) to obtain the fitting relationship: VC = VWC × k + l, with a coefficient of determination DC2, where k and l are constant parameters. S9 modeling steps: Randomly select half of the sampling points and extract the backscattering coefficient feature image pixel value sigma0 based on the coordinates. Quantitatively model the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and establish a quantitative model sigma0=F(UW) and sigma0=F(VWC). S10 Feature Image Inversion Steps: Substitute the backscattering coefficient feature image pixel values ​​corresponding to the remaining samples into the quantitative models sigma0=F(UW) and sigma0=F(VWC) to obtain the inversion values ​​of fresh weight per unit area UW' and water content per unit area VWC' of the remaining sampling points. Compare these values ​​with the measured values ​​to obtain the coefficient of determination DC3 between the inversion of fresh weight per unit area and the measured values, and the coefficient of determination DC4 between the measured and measured values ​​of water content per unit area. S11 Comparison and Screening Steps: Compare the sizes of DC1×DC3 and DC2×DC4. If DC1×DC3 > DC2×DC4, substitute each pixel value of the backscattering coefficient feature image into sigma0=F(UW) to calculate the fresh weight per unit area UW' of each pixel. Then, use the model VC=UW×γ to iterate through each pixel to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. Otherwise, substitute each pixel value of the backscattering coefficient feature image into the model sigma0=F(VWC) to calculate the water content per unit area VWC' of each pixel. Then, use the model VC=VWC×k+l to iterate through each pixel to obtain the fresh weight per unit area to obtain the 1-meter resolution radar-inverted rice carbon sink thematic map TIFF file of the experimental area. S12 Paddy Field Soil Elevation Measurement Steps: Extract the elevation value H of each sampling point from the 1-meter resolution digital elevation model, and simultaneously extract the average rice height h of the sampling points. Calculate the soil elevation H0=Hh of each sampling point, and use linear interpolation to construct a 1-meter resolution soil digital elevation model registered with the paddy field digital elevation model. Steps for creating the S13 rice height distribution thematic map: Difference the pixel value of each cell in the registered 1-meter resolution paddy field digital elevation model and the soil digital elevation model to obtain the TIFF file of the paddy field rice height distribution thematic map. S14 High-level carbon sequestration fitting steps: Fit the rice height h corresponding to the sampling point with the carbon sequestration per unit area VC: VC=p×h+q, where p and q are constant terms. S15 Step for height inversion carbon sink: The fitted model VC=p×h+q is used to traverse each pixel of the rice height distribution thematic map TIFF file to obtain the height inversion carbon sink thematic map. S16 Co-inversion Carbon Sequestration Steps: After registering the 1-meter resolution radar-inverted rice carbon sequestration thematic map TIFF file with the height-inverted carbon sequestration thematic image pixels, the two carbon sequestration values ​​of each pixel are added together and averaged to obtain the co-inverted rice carbon sequestration thematic map TIFF file.

2. The method for rice carbon sequestration retrieval based on UAV photogrammetry and Sentinel-1 satellite VV polarization coordination as described in claim 1, characterized in that, The process of obtaining the UAV-based digital elevation model of the paddy field in S1 specifically includes the following steps: S1-1 Rice Paddy Drone Digital 3D Model Creation: Data from the experimental field was collected by drone aerial photography, and a digital 3D model of the rice paddy was created using 2D Gaussian splashing technology. S1-2 converts the created 3D digital model of the paddy field into a 2D digital elevation model TIFF file.

3. The method for obtaining a digital elevation model of a paddy field using an unmanned aerial vehicle (UAV) according to claim 1, characterized in that, Step S6 specifically includes the following steps: Steps for measuring the unit area moisture content at sampling point S6-1: Calculate the average plant moisture content WC by subtracting the average fresh weight from the average dry weight; calculate the number of rice plants per unit area at each sampling point: N=n / (a×d); finally, obtain the unit area moisture content VWC=WC×N at the sampling point. S6-2 Sampling Point Unit Area Carbon Sequestration Measurement Steps: Unit Area Carbon Sequestration VC = C × N. Steps for calculating the fresh weight of rice per unit area at each sampling point in S6-3: Fresh weight of rice per unit area UW = W × N.

4. The method for obtaining a digital elevation model of a paddy field using an unmanned aerial vehicle (UAV) according to claim 1, characterized in that, Step S9 specifically includes the following steps: S9-1 Quantitative Fitting Steps: Extract the pixel value sigma0 of the backscattering coefficient feature image at 1-meter resolution and perform scatter analysis with the corresponding fresh weight per unit area UW and water content per unit area VWC of the sampling points, and then perform least squares fitting.

Images