Farmland soil exposure monitoring method and system based on time-series optical remote sensing images

Abstract

The present invention discloses a method and system for monitoring bare farmland soil based on time-series optical remote sensing images, obtains remote sensing images of farmland in a study area, calculates the optimal soil adjustment vegetation index threshold, and uses the threshold to judge the state of farmland pixels at each moment on the soil adjustment vegetation index change curve, so as to obtain a spatiotemporally continuous bare farmland soil monitoring result. The present invention effectively overcomes the observation discontinuity problem caused by satellite re-entry period and cloud coverage by fitting the SAVI value extracted from the time-series optical images, so as to accurately capture the dynamic changes of soil exposure. The present invention realizes the optimization of vegetation index thresholds for distinguishing bare soil from crops, ensures the rationality and applicability of the thresholds in the study area, and improves the accuracy of soil exposure monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of farmland soil monitoring, and in particular to a method and system for monitoring exposed farmland soil based on time-series optical remote sensing images. Background Art

[0002] Farmland soil is the cornerstone of crop growth, and its health plays a decisive role in the sustainability of agricultural production. However, some farmland operations, such as land preparation before sowing, straw treatment after harvest, fallow and abandonment, may lead to soil exposure. This soil exposure phenomenon increases the risk of soil erosion and soil loss, which in turn poses a potential threat to the stability of farmland ecosystems and the sustainable development of agricultural production. Traditional farmland soil health research is mostly focused on the site scale. These studies provide important scientific basis for clarifying the impact of farmland operations on soil characteristics, but their spatial coverage is limited, making it difficult to effectively monitor the dynamic changes in soil conditions over a large area.

[0003] In recent years, remote sensing technology has been widely used in the field of agricultural monitoring due to its significant advantages of large-scale and periodic observations. In particular, time-series optical remote sensing data, with its rich multi-temporal and multi-spectral information, can effectively capture the spatiotemporal dynamic changes of farmland status. However, there are still relatively few studies on the precise monitoring of large-scale soil exposure. These studies usually have a low monitoring frequency, and it is difficult to accurately capture the dynamic changes of soil exposure during frequent farmland operations and rapid crop growth stages. In addition, existing studies often use the vegetation index threshold method to extract exposed soil, but the setting of the threshold mainly relies on expert experience and lacks systematic algorithm support. How to design a reasonable threshold setting algorithm is an important problem to be solved in remote sensing monitoring of farmland soil exposure. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention proposes a method for monitoring farmland soil exposure based on time-series optical remote sensing images, which can realize spatiotemporal continuous monitoring of farmland soil exposure over a large area, and includes the following steps:

[0005] Step 1: Obtain time-series optical remote sensing image data of the study area and perform cloud removal processing;

[0006] Step 2, using land cover products to determine the extent of farmland within the study area;

[0007] Step 3, calculating the soil adjusted vegetation index using the reflectance data of multiple bands;

[0008] Step 4: With the assistance of high-resolution images, bare soil and crop pixels within the farmland range of the time-series optical remote sensing images of the study area are marked as samples, and the soil-adjusted vegetation index at the corresponding time is extracted;

[0009] Step 5, determine the optimal soil conditioned vegetation index threshold in the study area;

[0010] Step 6, fitting the effective soil-adjusted vegetation index data of each farmland pixel in the study area to obtain a continuous soil-adjusted vegetation index change curve;

[0011] Step 7, using the optimal soil conditioning vegetation index threshold determined in step 5, judge the state of the farmland pixels at each moment on the soil conditioning vegetation index change curve obtained in step 6, and obtain the spatiotemporally continuous farmland soil exposure monitoring results in the study area.

[0012] Preferably, in step 2, the farmland type in the land cover product that matches the spatial resolution of the time-series optical remote sensing image is selected as a mask to obtain the scope of farmland in the study area.

[0013] As a preference, the soil adjusted vegetation index in step 3 SAVI The calculation formula is:

[0014] (1)

[0015] In the formula, NIR , R Represent the reflectance of near infrared and red bands respectively; L Soil conditioning factor.

[0016] Preferably, in step 4, with the help of high-resolution images with higher resolution than the time-series optical remote sensing images, bare soil and crop pixels within the farmland range of the time-series optical remote sensing images are marked as samples, and the SAVI value of the corresponding time is calculated; due to the influence of the satellite re-entry cycle and cloud cover, there may be a situation where there is no available optical remote sensing image at the corresponding time. When this happens, a time interval of no more than The soil adjusted vegetation index of the pixel sample is calculated by using the optical remote sensing image of the day closest to the target time. Manually set as needed; if no image meeting the conditions is found, delete the sample.

[0017] Preferably, in step 5, the samples with SAVI values ​​greater than 99% quantile or less than 1% quantile in bare soil samples and crop samples are first deleted, and the initial range of the optimal threshold is determined according to the maximum SAVI value of the remaining bare soil samples and the minimum SAVI value of the crop samples. The percentile of the initial range is taken as the potential threshold for evaluation. When the SAVI value of the pixel is less than the potential threshold, it is judged as bare soil, and when the SAVI value of the pixel is greater than or equal to the potential threshold, it is judged as crop, and the evaluation score of each potential threshold is calculated according to the following formula: :

[0018] (2)

[0019] In the formula, Indicates k potential threshold, Indicates i The SAVI value of correctly classified samples, represents the number of correctly classified samples, Indicates j The SAVI value of misclassified samples, represents the number of misclassified samples, represents the sum of the distances between the potential threshold and the SAVI value of the correctly classified sample, represents the sum of the distances between the potential threshold and the SAVI value of the misclassified sample, It is a constant term, which is set to prevent the denominator of the formula from being zero due to the misclassified sample being zero.

[0020] The potential threshold with the largest evaluation score is selected as the optimal threshold.

[0021] Preferably, in step 6, the SAVI value extracted from each farmland pixel at the time when cloud coverage is excluded is used as the effective soil conditioned vegetation index data, and the effective SAVI value is fitted by time series harmonic analysis to obtain a continuous soil conditioned vegetation index change curve. The time series harmonic analysis calculation formula is:

[0022] (3)

[0023] In the formula, Represents pixels j In time t The reconstruction value of Indicates i The frequency of the harmonic components; , N is the total number of harmonic components; is a constant term; and Represents the coefficients of the triangular components.

[0024] Preferably, in step 7, the optimal soil adjusted vegetation index threshold determined in step 5 is used to judge whether the state of the farmland pixel on the SAVI curve obtained in step 6 at each moment is bare soil or crop, so as to obtain the spatiotemporally continuous farmland soil exposure monitoring results in the study area.

[0025] The present invention also provides a farmland soil exposure monitoring system based on time-series optical remote sensing images, which is used to implement the farmland soil exposure monitoring method based on time-series optical remote sensing images as described above.

[0026] Moreover, it includes a processor and a memory, the memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute the above-mentioned farmland soil exposure monitoring method based on time-series optical remote sensing images.

[0027] Alternatively, it includes a readable storage medium having a computer program stored thereon, and when the computer program is executed, the method for monitoring exposed farmland soil based on time-series optical remote sensing images as described above is implemented.

[0028] Compared with the prior art, the present invention has the following advantages:

[0029] 1) The present invention effectively overcomes the observation discontinuity problem caused by satellite re-entry period and cloud cover by fitting the SAVI values ​​extracted from time-series optical images, so as to accurately capture the dynamic changes of soil exposure; 2) The present invention realizes the optimization of vegetation index thresholds for distinguishing bare soil from crops, ensures the rationality and applicability of the thresholds in the study area, and improves the accuracy of soil exposure monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 It is a flow chart of farmland soil exposure monitoring based on time-series optical remote sensing images according to an embodiment of the present invention.

[0032] Figure 2 is a schematic diagram of a preferred vegetation index threshold for distinguishing between bare soil and crops in an embodiment of the present invention. Figure 2 (1) in the middle indicates screening of crop samples with SAVI values ​​in the 1%-99% percentile range. Figure 2 (2) in the middle indicates the selection of bare soil samples with SAVI values ​​in the 1%-99% quantile range. Figure 2 (3) indicates determining the optimal initial range of the threshold. Figure 2 (4) indicates determining the optimal threshold based on the evaluation score. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in combination with the drawings and 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.

[0034] Example 1

[0035] like Figure 1 As shown, an embodiment of the present invention provides a method for monitoring exposed farmland soil based on time-series optical remote sensing images, comprising the following steps:

[0036] Step 1: Obtain the time-series optical remote sensing image data of the study area and perform cloud removal processing.

[0037] Acquire time-series Sentinel-2 image data of the study area, use the Sentinel-2 Cloud Probability product to remove clouds from Sentinel-2 images, and remove pixels with cloud probability exceeding 50%.

[0038] Step 2: Use land cover products to determine the extent of farmland in the study area.

[0039] The cropland types in the WorldCover 2020 land cover product released by the European Space Agency, which matches the spatial resolution of the Sentinel-2 image, were selected as masks to obtain the extent of cropland in the study area.

[0040] Step 3, using the reflectance data of multiple bands to calculate the soil adjusted vegetation index, which is used to distinguish between bare soil and crops within the farmland.

[0041] The soil-adjusted vegetation index (SVI) is used, which is more sensitive to changes in land cover under sparse vegetation conditions. SAVI ) to distinguish between bare soil and crops within the farmland, the calculation formula of SAVI is:

[0042] (1)

[0043] In the formula, NIR , R Represent the reflectance of near infrared and red bands respectively; L is the soil adjustment factor, and in this embodiment, the value is 0.5.

[0044] Step 4: With the assistance of high-resolution images, bare soil and crop pixels within the farmland range of the time-series optical remote sensing images of the study area are marked as samples, and the soil-adjusted vegetation index at the corresponding time is extracted.

[0045] Using high-resolution images in Google Earth Pro, bare soil and crop pixels within the farmland range of the time series Sentinel-2 images were annotated as samples, and the corresponding time was calculated. SAVI Due to the influence of satellite re-entry period and cloud cover, there may be no available Sentinel-2 image at the corresponding time. When this happens, the Sentinel-2 image with a time interval of no more than 3 days and closest to the target time is used to calculate the soil adjusted vegetation index of the pixel sample; if no image that meets the conditions can be found, the sample is deleted.

[0046] Step 5, determine the optimal soil conditioned vegetation index threshold in the study area to minimize the probability of misclassification of bare soil and crop samples.

[0047] like Figure 2 As shown in (1) and (2), to avoid the influence of outliers on threshold selection, the samples with SAVI values ​​greater than the 99% quantile or less than the 1% quantile in bare soil samples and crop samples are first deleted. The initial range of the optimal threshold is determined based on the largest SAVI value of the remaining bare soil samples and the smallest SAVI value of the crop samples ( Figure 2 In (3), the percentile of the initial range is taken as the potential threshold for evaluation. When the SAVI value of a pixel is less than the potential threshold, it is judged as bare soil. When the SAVI value of a pixel is greater than or equal to the potential threshold, it is judged as crop. The evaluation score of each potential threshold is calculated according to the following formula: :

[0048] (2)

[0049] In the formula, Indicates k potential threshold, Indicates i The SAVI value of correctly classified samples, represents the number of correctly classified samples, Indicates j The SAVI value of misclassified samples, represents the number of misclassified samples, represents the sum of the distances between the potential threshold and the SAVI value of the correctly classified sample, represents the sum of the distances between the potential threshold and the SAVI value of the misclassified sample, is a constant term, which is set to prevent the denominator of the formula from being zero due to the misclassified sample being zero. In this embodiment, the value is 1.

[0050] Since the optimal threshold should make the correctly classified samples as far away from the threshold as possible to ensure separability, and make the incorrectly classified samples as close to the threshold as possible to reduce the probability of misclassification, the potential threshold with the largest evaluation score is finally selected as the optimal threshold, which is 0.1864 ( Figure 2 in (4).

[0051] Step 6: Fit the effective soil-adjusted vegetation index data of each farmland pixel in the study area to obtain a continuous soil-adjusted vegetation index change curve.

[0052] The SAVI value extracted from each farmland pixel at the time when cloud coverage is excluded is used as the effective soil conditioned vegetation index data. The time series harmonic analysis method (HANTS) is used to fit the effective SAVI value to obtain a continuous soil conditioned vegetation index change curve. The time series harmonic analysis calculation formula is:

[0053] (3)

[0054] In the formula, Represents pixels j In time t The reconstruction value of Indicates i The frequency of the harmonic components; , N is the total number of harmonic components, which is 4 in this embodiment; is a constant term; and Represents the coefficients of the triangular components.

[0055] Step 7, using the determined optimal soil adjusted vegetation index threshold to judge the state of farmland pixels at each moment, to obtain a spatiotemporally continuous farmland soil exposure monitoring result.

[0056] Using the optimal soil adjusted vegetation index threshold determined in step 5, the state of the farmland pixel on the SAVI curve obtained in step 6 at each moment is judged to be bare soil or crop, and the spatiotemporally continuous farmland soil exposure monitoring results in the study area are obtained.

[0057] Example 2

[0058] Based on the same inventive concept, the present invention also provides a farmland soil exposure monitoring system based on time-series optical remote sensing images, including a processor and a memory, the memory is used to store program instructions, and the processor is used to call the program instructions in the memory to execute a farmland soil exposure monitoring method based on time-series optical remote sensing images as described above.

[0059] Example 3

[0060] Based on the same inventive concept, the present invention also provides a farmland soil exposure monitoring system based on time-series optical remote sensing images, including a readable storage medium, on which a computer program is stored. When the computer program is executed, a farmland soil exposure monitoring method based on time-series optical remote sensing images as described above is implemented.

[0061] In specific implementation, the method proposed in the technical solution of the present invention can be implemented by technical personnel in this field using computer software technology to realize automatic operation process. System devices for implementing the method, such as computer-readable storage media storing the corresponding computer program of the technical solution of the present invention and computer equipment running the corresponding computer program, should also be within the scope of protection of the present invention.

[0062] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific implementation cases described, or replace them with similar methods, without departing from the spirit of the present invention or exceeding the scope defined by the appended claims.

Claims

1. A method for monitoring exposed farmland soil based on time-series optical remote sensing images, characterized in that: The following steps are involved: Step 1: Obtain time-series optical remote sensing image data of the study area and perform cloud removal processing; Step 2, using land cover products to determine the extent of farmland within the study area; Step 3, calculating the soil adjusted vegetation index using the reflectance data of multiple bands; Step 4: With the assistance of high-resolution images, bare soil and crop pixels within the farmland range of the time-series optical remote sensing images of the study area are marked as samples, and the soil-adjusted vegetation index at the corresponding time is extracted; Step 5, determine the optimal soil conditioned vegetation index threshold in the study area; First, the samples with SAVI values ​​greater than 99% quantile or less than 1% quantile in bare soil samples and crop samples were deleted respectively. The initial range of the optimal threshold was determined according to the largest SAVI value of the remaining bare soil samples and the smallest SAVI value of the crop samples. The percentile of the initial range was taken as the potential threshold for evaluation. When the SAVI value of the pixel was less than the potential threshold, it was judged as bare soil. When the SAVI value of the pixel was greater than or equal to the potential threshold, it was judged as crop. The evaluation score of each potential threshold was calculated according to the following formula: : （2） In the formula, Indicates k potential threshold, Indicates i The SAVI value of correctly classified samples, represents the number of correctly classified samples, Indicates j The SAVI value of misclassified samples, represents the number of misclassified samples, represents the sum of the distances between the potential threshold and the SAVI value of the correctly classified sample, represents the sum of the distances between the potential threshold and the SAVI value of the misclassified sample, It is a constant term, which is set to prevent the denominator of the formula from being zero due to the misclassified sample being zero; The potential threshold with the largest evaluation score is selected as the optimal threshold; Step 6, fitting the effective soil-adjusted vegetation index data of each farmland pixel in the study area to obtain a continuous soil-adjusted vegetation index change curve; Step 7, using the optimal soil conditioning vegetation index threshold determined in step 5, judge the state of the farmland pixels at each moment on the soil conditioning vegetation index change curve obtained in step 6, and obtain the spatiotemporally continuous farmland soil exposure monitoring results in the study area.

2. The method for monitoring exposed farmland soil based on time-series optical remote sensing images according to claim 1, characterized in that: In step 2, the farmland type in the land cover product that matches the spatial resolution of the time-series optical remote sensing image is selected as a mask to obtain the scope of farmland in the study area.

3. The method for monitoring exposed farmland soil based on time-series optical remote sensing images according to claim 1, characterized in that: Soil Adjusted Vegetation Index in Step 3 SAVI The calculation formula is: （1） In the formula, NIR , R Represent the reflectance of near infrared and red bands respectively; L Soil conditioning factor.

4. The method for monitoring exposed farmland soil based on time-series optical remote sensing images according to claim 1, characterized in that: In step 4, with the help of high-resolution images with higher resolution than the time-series optical remote sensing images, bare soil and crop pixels within the farmland range of the time-series optical remote sensing images are marked as samples, and the corresponding time is calculated. SAVI value; Due to the influence of satellite re-entry period and cloud cover, there may be no available optical remote sensing images at the corresponding time. When this happens, the time interval should not exceed The soil adjusted vegetation index of the pixel sample is calculated by using the optical remote sensing image of the day closest to the target time. Manually set as needed; if no image meeting the conditions is found, delete the sample.

5. The method for monitoring exposed farmland soil based on time-series optical remote sensing images according to claim 1, characterized in that: In step 6, the SAVI value extracted from each farmland pixel at the time when cloud coverage is excluded is used as the effective soil conditioned vegetation index data. The effective SAVI value is fitted using the time series harmonic analysis method to obtain a continuous soil conditioned vegetation index change curve. The time series harmonic analysis calculation formula is: （3） In the formula, Represents pixels j In time t The reconstruction value of Indicates i The frequency of the harmonic components; , N is the total number of harmonic components; is a constant term; and Represents the coefficients of the triangular components.

6. The method for monitoring exposed farmland soil based on time-series optical remote sensing images according to claim 1, characterized in that: In step 7, the optimal soil adjusted vegetation index threshold determined in step 5 is used to determine whether the state of the farmland pixels on the SAVI curve obtained in step 6 at each moment is bare soil or crop, and the spatiotemporally continuous farmland soil exposure monitoring results in the study area are obtained.

7. A farmland soil exposure monitoring system based on time-series optical remote sensing images, characterized in that: It includes a processor and a memory, the memory is used to store program instructions, and the processor is used to call the program instructions in the memory to execute a farmland soil exposure monitoring method based on time-series optical remote sensing images as described in any one of claims 1-6.

8. A farmland soil exposure monitoring system based on time-series optical remote sensing images, characterized in that: It includes a readable storage medium, on which a computer program is stored. When the computer program is executed, a method for monitoring exposed farmland soil based on time-series optical remote sensing images as described in any one of claims 1 to 6 is implemented.

Images