Farmland crop irrigation water consumption estimation method based on remote sensing soil moisture data

Abstract

The invention discloses a farmland crop irrigation water consumption estimation method based on remote sensing soil moisture data, and belongs to the technical field of agricultural irrigation. The method comprises the following steps: S1, determining a farmland range to be estimated; s2, acquiring and preprocessing data required for estimation corresponding to the farmland range, wherein the data required for estimation comprises a soil moisture remote sensing image and rainfall data; s3, extracting a soil moisture time sequence of each pixel based on the soil moisture remote sensing image; s4, smoothing the soil moisture time sequence; s5, determining an irrigation time period based on the smoothed soil moisture time sequence; and S6, estimating the total irrigation amount in a specific time length according to the soil moisture time sequence, the irrigation time period and the rainfall data. According to the invention, high-precision monitoring of crop raw season irrigation information in a large-area field range can be realized, and the problems of more input data and complex method model existing in the prior art of using a water balance model to quantify irrigation are solved.

Description

Technical Field

[0001] The invention belongs to the technical field of agricultural irrigation, and in particular relates to a method for estimating the amount of water used for irrigation of farmland crops based on remote sensing soil moisture data. Background Art

[0002] Irrigation is one of the largest human interventions in the water cycle, using more than 70% of global freshwater withdrawals. Today, about 20% of the world's arable land is irrigated, which provides more than 40% of the world's food. Understanding irrigation information is critical to preventing water waste, avoiding illegal water withdrawals, and ensuring food and water security.

[0003] However, most current irrigation reports based on statistical surveys can only provide the total number of irrigated areas and total irrigation water in a specific region, and they are spatially inconsistent, have low resolution, and large errors, such as the China Water Resources Bulletin and the water use report issued every five years by the United States Geological Survey. Therefore, it is still difficult to obtain the spatial distribution of each irrigated farmland and its irrigation water. Most studies focus on mapping irrigation using remote sensing data, while the extraction of irrigation time and irrigation amount remains challenging. Typically, some studies use actual evapotranspiration (ETa), remote sensing estimates (such as MODIS and Landsat), and water balance modeling methods to indirectly assess irrigation water. However, on the one hand, the amount of input data is large and the method is complex, and on the other hand, it is limited by weather conditions. In addition, there is a lack of field observations of irrigation water use, so it is difficult to assess the reliability and accuracy of the resulting estimates. It is well known that soil moisture is the most direct indicator for inferring irrigation information. Once irrigation is carried out, soil moisture may continue to increase until it reaches a stable state, and this irrigation signal can be detected by in situ or satellite sensors. Therefore, methods to directly use high-resolution soil moisture time series to extract pixel-scale irrigation information are urgently needed. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the above-mentioned related art at least to a certain extent.

[0005] To this end, the purpose of the present invention is to provide a method for estimating farmland crop irrigation water consumption based on remote sensing soil moisture data, which can achieve high-precision monitoring of crop irrigation information during the growing season in a large field area, and solve the problems of the existing use of water balance models to quantify irrigation, such as the large amount of input data and complex method models.

[0006] In order to solve the above-mentioned technical problems, the present invention is achieved as follows: An embodiment of the present invention provides a method for estimating water consumption for crop irrigation based on remote sensing soil moisture data, the method comprising: S1. Determine the scope of farmland to be estimated; S2, obtaining and preprocessing the data required for estimation corresponding to the farmland range, the data required for estimation including soil moisture remote sensing images and precipitation data; S3, extracting soil moisture time series of each pixel based on soil moisture remote sensing image; S4, smoothing the soil moisture time series; S5, determining the irrigation period based on the smoothed soil moisture time series; S6. Estimate the total irrigation volume for a specific period of time based on soil moisture time series, irrigation period, and precipitation data.

[0007] In addition, the method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to the present invention may also have the following additional technical features: In some embodiments, the method further comprises: S7. Calculate a root mean square error based on the estimated total irrigation amount and the actual irrigation amount, and determine the accuracy of the estimated total irrigation amount based on the root mean square error.

[0008] In some of the embodiments, the preprocessing in step S2 includes: cropping the remote sensing image in the spatial dimension and then splicing it in the time dimension to obtain a series of remote sensing images of a single pixel in the time dimension, thereby obtaining the soil moisture changes of each pixel in the farmland in the time dimension.

[0009] In some implementations, the smoothing algorithm in step S4 is a moving average smoothing method, exponential smoothing or a Savitzky-Golay filtering algorithm.

[0010] In some implementations, in step S4, an exponential smoothing method is used to smooth the soil moisture time series.

[0011] In some of the embodiments, the content of smoothing using the exponential smoothing method includes: inputting the soil moisture time series of each pixel within the farmland range, adjusting the seasonal cycle length parameter of the soil moisture time series, and obtaining the optimal smoothing parameter for the purpose of retaining the main characteristics of the soil moisture curve.

[0012] In some of the embodiments, the content of step S5 includes: traversing the soil moisture time series data of each pixel, identifying each process of continuous increase of soil moisture and the soil moisture difference of each process; Set a reasonable threshold and compare each soil moisture difference with the threshold; If the soil moisture difference exceeds the threshold, the point where the soil moisture begins to increase during the process is marked as the starting point, and the point where the soil moisture difference no longer increases or is about to show a downward trend is marked as the end point, thereby determining a complete irrigation period; The thresholds of each identified process of continuous increase in soil moisture are compared and judged to obtain all irrigation periods.

[0013] In some implementations, the root mean square error is calculated as: (2) in, is the actual irrigation volume collected, is the estimated irrigation volume, n Indicates the number of irrigation times.

[0014] In some embodiments, the total irrigation amount of the study area in step S6 is calculated as follows: (1)

[0015] in, Q is the irrigation volume of crops in the study area during the growing season, θ i+1 - θ i For a pixel i The difference between the maximum and minimum soil moisture in each irrigation period. h is the soil depth in m, m is the number of pixels in the study area, n Indicates the number of irrigation times. P is the total precipitation in the study area, in mm.

[0016] In some of these embodiments, the threshold is 5%.

[0017] Compared with the prior art, the present invention has at least the following beneficial effects: In an embodiment of the present invention, a method for estimating the amount of water used for irrigation of farmland crops based on remote sensing soil moisture data is provided, which achieves the purpose of directly extracting the field irrigation amount using the soil moisture time series at the pixel scale, and is expected to solve the problems in the current method of having many types of data inputs, complex methods, and only being able to provide irrigation water consumption but not the actual irrigation amount of field crops.

[0018] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1A flowchart of a method for estimating water consumption for crop irrigation based on remote sensing soil moisture data disclosed in one embodiment of the present invention; Figure 2 This is an irrigation period diagram in a certain pixel soil moisture time series disclosed in an embodiment of the present invention, wherein the yellow point is the starting point of irrigation and the green line is the end point of irrigation, and the blue arrow is the difference between the maximum and minimum values ​​of the irrigation period. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments 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.

[0021] The embodiments of the present invention are described in detail below through specific embodiments and application scenarios in conjunction with the accompanying drawings.

[0022] See also Figure 1 As shown, in some embodiments of the present invention, a method for estimating the amount of irrigation water for farmland crops based on remote sensing soil moisture data is provided. The method directly uses the soil moisture time series at the pixel scale to extract field irrigation amount data, and the steps include: Step 1: Data Collection Collect soil moisture remote sensing images with high spatiotemporal resolution (30 meters, every day) in the selected study area. These images will serve as the basic data source for subsequent analysis. Based on remote sensing data processing technology, such as using remote sensing data processing software such as ENVI, perform operations such as cropping and splicing on the high-spatial remote sensing images of the study area to ensure that the soil moisture changes of each pixel in the study area in the time dimension can be obtained. That is, the same spatial remote sensing image is cropped and spliced ​​according to time to obtain a series of remote sensing images of a single pixel in the time dimension, and then the soil moisture changes of the pixel can be analyzed. This application does not explicitly limit the data source, mosaic method, etc.

[0023] At the same time, comprehensive precipitation data in the study area will be collected, which will help analyze the impact of precipitation on soil moisture and its subsequent relationship with irrigation volume. In addition, actual irrigation volume data in the study area needs to be collected. This information comes from field irrigation records, agricultural department statistics, etc. These actual measurement data are crucial for calibrating the model and accurately evaluating the irrigation situation.

[0024] Step 2: Extract farmland range Using land type data with the same resolution as remote sensing images, with the help of professional geographic information system (GIS) software and data analysis tools, the farmland scope of the study area is extracted and determined. Through detailed interpretation and analysis of land type data, different land categories can be distinguished and farmland areas can be accurately located. The order of step 2 and step 1 can be swapped to ensure that the farmland scope used in the end corresponds to the data obtained in step 1.

[0025] Step 3: Extract soil moisture time series If the long-term soil moisture remote sensing products of the farmland in the study area can be directly obtained, the soil moisture values ​​of each pixel in the time dimension corresponding to each image can be directly extracted and superimposed to obtain the soil moisture time series. If the above long-term soil moisture remote sensing products cannot be directly obtained, it is necessary to download the remote sensing images for inversion and then extract the soil moisture time series. The steps include: downloading the remote sensing images corresponding to the study area (such as Landsat satellite) on Google Earth Engine and preprocessing them at the same time, including radiation calibration, atmospheric correction, mosaic, cropping, etc.; then, based on the visible light / near infrared and thermal infrared spatial information inversion method, the soil moisture distribution image is obtained; specifically, the surface temperature (LST) and vegetation index (NDVI) are obtained through the processed remote sensing images to form the temperature-vegetation index feature space to determine the dry edge and wet edge, and the soil moisture volume content of the study area is further obtained through subsequent calculations; then the soil moisture values ​​of each pixel corresponding to each image are extracted and superimposed to synthesize the soil moisture time series.

[0026] Step 4: Smoothing the soil moisture time series This step aims to smooth the soil moisture time series of each pixel to reduce the interference of noise on subsequent analysis while retaining the key features of the curve to the greatest extent. Select a suitable smoothing algorithm to smooth the time series, such as moving average smoothing, exponential smoothing or Savitzky-Golay filtering algorithm. Through multiple experiments and parameter adjustments, find an optimal smoothing scheme that can effectively reduce the noise level and clearly display the main features of the soil moisture curve (such as mutations caused by irrigation or precipitation events), thereby laying a solid data foundation for the subsequent accurate extraction of irrigation characteristics.

[0027] In some embodiments of the present invention, the extracted soil moisture time series data is smoothed using an exponential smoothing method, which is a commonly used time series data smoothing method that can be used to process data with trends and seasonality, and can handle trends, seasonality and noise well. In this process, the soil moisture time series of each pixel in the study area is first input, and then the seasonal cycle length parameter of the soil moisture time series is adjusted to retain the best smoothing scheme in which the main features of the soil moisture curve are clearly displayed, thereby laying a solid data foundation for the subsequent accurate extraction of irrigation characteristics.

[0028] In the above implementation, in the process of smoothing the soil moisture time series, the relationship between the degree of smoothing and the retention of curve features needs to be carefully weighed. If the smooth transition is too smooth, the true trend and key features of soil moisture changes may be blurred, such as the rapid rise caused by irrigation or the slow decline caused by drought; if the smoothing is insufficient, the noise interference cannot be effectively removed.

[0029] Step 5: Determine the irrigation period The core of this step is to accurately identify the characteristic time periods related to irrigation activities from the smoothed soil moisture time series data. In the specific operation, a reasonable threshold is set, the soil moisture time series data of each pixel is traversed, and the soil moisture difference is calculated. When the difference exceeds the set threshold, the beginning of the increase is marked as the starting point; continue to search backward until the soil moisture difference no longer increases or shows a clear downward trend, and mark it as the end point of the period, thereby determining a complete possible irrigation period.

[0030] In the above implementation, the threshold is set to 5%. If the difference between the beginning and the end of the soil moisture increase is greater than 5%, the period of continuous soil moisture increase is identified. Figure 2 As shown in the figure, when the soil moisture increases from θ1 to θ2 and its increase exceeds 5%, it is determined as an irrigation period, and the moment when the soil moisture begins to increase is marked as the starting point (marked by a yellow dot). Then continue to search backward until the turning point before the soil moisture value slowly decreases or significantly decreases due to the absorption or evaporation of irrigation water, that is, after the soil moisture value reaches the highest point, the value of the next day is continuously smaller than the value of the previous day. At this time, this point is marked as the end point of the irrigation period (marked by a green dot).

[0031] In some embodiments of the present application, the threshold value of this step needs to be appropriately adjusted according to factors such as soil type, climate conditions, and crop growth characteristics of the study area. For example, in sandy soil areas, due to the poor water retention of the soil, the increase in soil moisture after irrigation may be relatively small, and the threshold value may need to be appropriately lowered; in areas with frequent precipitation, it is necessary to be more cautious in distinguishing the effects of irrigation and precipitation on soil moisture, and it may be necessary to combine precipitation data and the time series characteristics of soil moisture changes to further optimize the selection of thresholds to ensure accurate extraction of irrigation feature periods.

[0032] Step 6: Calculate irrigation volume In the time period when each pixel has irrigation characteristics determined in the previous step, the maximum and minimum values ​​of soil moisture in each time period are subtracted and the differences are accumulated to determine the net increase in soil moisture of the pixel in a longer period of time (including multiple irrigation periods). At the same time, the contribution of precipitation to soil moisture in the study area is subtracted from the net increase in soil moisture of all pixels, and finally the total irrigation amount of the entire study area in a specific growing season is obtained.

[0033] In some embodiments of the present invention, the calculation formula for the total irrigation amount of the study area in this step is as follows: (1) in, Q is the irrigation volume of crops in the study area during the growing season, in mm, θ i+1 - θ i It is the difference between the maximum and minimum soil moisture in a certain irrigation period within a pixel, in m 3 / m 3 , h is the soil depth in m, m is the number of pixels in the study area, P is the total precipitation in the study area, in mm.

[0034] In some embodiments of the present invention, first, for each pixel, the soil moisture data within the irrigation period is determined, and the difference between the maximum value and the minimum value is obtained, corresponding to Figure 2 The difference between the green point and the yellow point in one irrigation stage. For example, θ2-θ1 in the figure is the soil moisture difference of one irrigation. According to formula (1), the difference of the irrigation period of all pixels in the study area is accumulated, and the volumetric water content of soil moisture is converted into equivalent depth. This accumulation result reflects to a certain extent the overall increase in soil moisture in the study area when the influence of precipitation is not considered. At the same time, the total precipitation in the study area obtained in the same period is removed to obtain the total irrigation amount of the study area.

[0035] Step 7: Accuracy Verification Combined with the actual irrigation information, soil characteristics, crop water demand patterns and other relevant information of the study area, the calculation results in the previous steps are rationally verified and necessary corrections are made to the actual irrigation volume to improve the accuracy and reliability of the irrigation volume calculation.

[0036] In some embodiments of the present invention, the desired irrigation amount is compared with the actual irrigation amount, and the accuracy of the desired irrigation amount is determined by the root mean square error formula (RMSE). The formula is as follows: (2) in, is the actual irrigation volume collected, is the estimated irrigation volume, n Indicates the number of irrigation times.

[0037] The smaller the RMSE value, the smaller the difference between the desired irrigation amount and the actual irrigation amount, and the higher the calculation accuracy; conversely, the larger the RMSE value, the lower the accuracy. By calculating RMSE, the accuracy of the desired irrigation amount can be quantitatively evaluated, providing a more scientific and accurate data basis for decision-making and application in related fields.

[0038] Parts not described in detail in the present invention may refer to the prior art in the art or are known to those skilled in the art, and the present invention will not elaborate on them in detail.

[0039] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data, characterized in that: The method comprises: S1. Determine the scope of farmland to be estimated; S2, obtaining and preprocessing the data required for estimation corresponding to the farmland range, the data required for estimation including soil moisture remote sensing images and precipitation data; S3, extracting soil moisture time series of each pixel based on soil moisture remote sensing image; S4, smoothing the soil moisture time series; S5, determining the irrigation period based on the smoothed soil moisture time series; S6. Estimate the total irrigation volume for a specific period of time based on soil moisture time series, irrigation period, and precipitation data.

2. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: The method further comprises: S7. Calculate a root mean square error based on the estimated total irrigation amount and the actual irrigation amount, and determine the accuracy of the estimated total irrigation amount based on the root mean square error.

3. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: The preprocessing in step S2 includes: cropping the remote sensing image in the spatial dimension and then splicing it in the temporal dimension to obtain a series of remote sensing images of a single pixel in the temporal dimension, thereby obtaining the soil moisture changes of each pixel in the farmland in the temporal dimension.

4. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: The smoothing algorithm in step S4 is a moving average smoothing method, exponential smoothing or a Savitzky-Golay filtering algorithm.

5. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: In step S4, the soil moisture time series is smoothed using an exponential smoothing method.

6. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 5, characterized in that: The content of smoothing by exponential smoothing method includes: inputting soil moisture time series of each pixel within the farmland range, adjusting the seasonal cycle length parameter of soil moisture time series, and obtaining the optimal smoothing parameter with the purpose of retaining the main characteristics of soil moisture curve.

7. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: The content of step S5 includes: traversing the soil moisture time series data of each pixel, identifying each process of continuous increase of soil moisture and the soil moisture difference of each process; Set a reasonable threshold and compare each soil moisture difference with the threshold; If the soil moisture difference exceeds the threshold, the point where the soil moisture begins to increase during the process is marked as the starting point, and the point where the soil moisture difference no longer increases or is about to show a downward trend is marked as the end point, thereby determining a complete irrigation period; The thresholds of each identified process of continuous increase in soil moisture are compared and judged to obtain all irrigation periods.

8. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 2, characterized in that: The formula for calculating the root mean square error is: （2） in, is the actual irrigation volume collected, is the estimated irrigation volume, n Indicates the number of irrigation times.

9. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 1, characterized in that: The calculation formula for the total irrigation amount of the study area in step S6 is as follows: （1） in, Q is the irrigation volume of crops in the study area during the growing season, θ i+1 - θ i For a pixel i The difference between the maximum and minimum soil moisture in each irrigation period. h is the soil depth in m, m is the number of pixels in the study area, n Indicates the number of irrigation times. P is the total precipitation in the study area, in mm.

10. The method for estimating the amount of water used for crop irrigation based on remote sensing soil moisture data according to claim 7, characterized in that: The threshold is 5%.

Images