A rice extraction method and device based on multi-temporal SAR data

Abstract

This application discloses a rice extraction method and device based on multi-temporal SAR data, relating to the field of agricultural production monitoring technology, with the primary purpose of improving rice identification and extraction accuracy. The main technical solution is as follows: obtaining multi-temporal SAR data corresponding to the target year for the area to be extracted and the corresponding backscatter coefficients; synthesizing the multi-temporal SAR data into a multi-band SAR image according to preset backscatter coefficient combinations for different temporal phases; determining the backscatter coefficient combination classification thresholds corresponding to each band based on the backscatter coefficient combination values ​​corresponding to each band in a second multi-band SAR image from a historical year and the proportion of rice distribution; introducing the backscatter coefficient combination classification thresholds corresponding to each band into a preset extraction rule, and using the extraction rule to extract rice from the first multi-band SAR image of the current year, thereby obtaining a rice distribution map corresponding to the area to be extracted in the current year.

Description

Technical Field

[0001] The present application relates to the technical field of agricultural production monitoring, and in particular to a rice extraction method and device based on multi-temporal SAR data. Background Art

[0002] With global climate change and population growth, food security is receiving increasing attention. As one of the world's major food crops, rice's cultivated area and yield have a significant impact on global food supply. To improve the efficiency and sustainability of agricultural production, accurately monitoring rice planting distribution, growth status, and yield prediction are crucial. With the continuous advancement of remote sensing technology, Synthetic Aperture Radar (SAR) data has become an effective means of monitoring and extracting information about crops like rice, due to its unaffected weather conditions and ability to be acquired around the clock.

[0003] Currently, existing techniques for extracting rice from Synthetic Aperture Radar (SAR) data typically rely solely on the polarization value of the backscatter coefficient. However, rice has different physical structures and growth states at different growth stages. Relying on a single polarization value cannot fully reflect the complex changes in rice throughout its various growth stages, resulting in low rice identification and extraction accuracy. Summary of the Invention

[0004] In view of the above problems, the present application provides a rice extraction method and device based on multi-temporal SAR data, the main purpose of which is to improve the accuracy of rice identification and extraction.

[0005] To solve the above technical problems, this application proposes the following solutions:

[0006] In a first aspect, the present application provides a rice extraction method based on multi-temporal SAR data, the method comprising:

[0007] Acquire multi-temporal SAR data corresponding to a target year for the area to be extracted and a backscatter coefficient corresponding to the multi-temporal SAR data, wherein the target year includes a current year and a historical year;

[0008] synthesizing the multi-temporal SAR data into a multi-band SAR image according to a combination of backscatter coefficients preset for different temporal phases, wherein the multi-band SAR image includes a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year;

[0009] Determining a backscatter coefficient combination classification threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution;

[0010] The backscatter coefficient combination classification threshold corresponding to each band is introduced into a preset extraction rule, and the extraction rule is used to extract rice from the first multi-band SAR image of the current year to obtain a rice distribution map corresponding to the area to be extracted in the current year.

[0011] In a second aspect, the present application provides a rice extraction device based on multi-temporal SAR data, the device comprising:

[0012] An acquisition unit, configured to acquire multi-temporal SAR data corresponding to a target year of an area to be extracted and a backscatter coefficient corresponding to the multi-temporal SAR data, wherein the target year includes a current year and a historical year;

[0013] a processing unit configured to synthesize the multi-temporal SAR data obtained by the acquisition unit into a multi-band SAR image according to a combination of backscatter coefficients preset for different temporal phases, wherein the multi-band SAR image includes a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year;

[0014] a first determining unit, configured to determine a backscatter coefficient combination classification threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year obtained by the processing unit and the proportion of rice distribution;

[0015] an extraction unit, configured to introduce the backscatter coefficient combination classification threshold corresponding to each band obtained by the first determination unit into a preset extraction rule, and use the extraction rule to perform rice extraction on the first multi-band SAR image of the current year obtained by the processing unit, to obtain a rice distribution map corresponding to the area to be extracted in the current year.

[0016] In order to achieve the above-mentioned purpose, according to the third aspect of the present application, a storage medium is provided, which includes a stored program, wherein when the program is run, the device where the storage medium is located is controlled to execute the rice extraction method based on multi-temporal SAR data according to the first aspect.

[0017] In order to achieve the above-mentioned object, according to a fourth aspect of the present application, a processor is provided, which is used to run a program, wherein when the program is run, the rice extraction method based on multi-temporal SAR data of the above-mentioned first aspect is executed.

[0018] By means of the above technical solution, the present application provides a rice extraction method and device based on multi-temporal SAR data. When it is necessary to extract the rice distribution, the multi-temporal SAR data corresponding to the current year and the historical year of the area to be extracted are first obtained. Then, according to the preset backscatter coefficient combination method of different phases, the multi-temporal SAR data are synthesized into a multi-band SAR image to obtain a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year. Then, according to the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution, the backscatter coefficient combination classification threshold corresponding to each band is determined. Finally, the backscatter coefficient combination classification threshold corresponding to each band is introduced into the preset extraction rule, and the extraction rule is used to extract rice from the first multi-band SAR image of the current year to obtain a rice distribution map corresponding to the area to be extracted in the current year. The technical solution provided in this application introduces multi-temporal SAR data of the current year and historical years, and synthesizes the multi-temporal SAR data according to the combination of backscatter coefficients of different phases, so that the obtained multi-band SAR image can more comprehensively reflect the scattering characteristics of rice in different years and different growth stages, so that it can capture more details and features, which helps to more accurately identify and extract rice. The multi-band SAR images of historical years are used to determine the rice extraction threshold corresponding to the multi-band SAR image of the current year. Past experience and rules are fully considered, providing a reliable reference standard for rice extraction in the current year, reducing uncertainty and error, and accurately reflecting the rice distribution in the area to be extracted in the current year, thereby effectively improving the accuracy of rice identification and extraction.

[0019] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0021] Figure 1 A flow chart of a rice extraction method based on multi-temporal SAR data provided in an embodiment of the present application is shown;

[0022] Figure 2 A flow chart of another rice extraction method based on multi-temporal SAR data provided in an embodiment of the present application is shown;

[0023] Figure 3 A block diagram of a rice extraction device based on multi-temporal SAR data provided by an embodiment of the present application is shown;

[0024] Figure 4 A block diagram of another rice extraction device based on multi-temporal SAR data provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0026] Currently, existing techniques for extracting rice from Synthetic Aperture Radar (SAR) data typically rely solely on polarization values ​​within the backscatter coefficient, such as the horizontal polarization value (VV) or the vertical polarization value (VH). However, rice has different physical structures and growth states during different growth periods. Relying on a single polarization value cannot fully reflect the complex changes in rice at various growth stages, nor does it fully incorporate the combined characteristics of the backscatter coefficients throughout the rice growth cycle, resulting in low rice identification and extraction accuracy.

[0027] After research, the inventors discovered that it is possible to integrate multi-temporal SAR data from historical and current years and synthesize them into multi-band SAR images based on preset backscatter coefficient combinations for different phases to cover the changing characteristics of rice throughout its entire growth cycle. Furthermore, based on the backscatter coefficient combination values ​​of each band in the multi-band SAR images from historical years and the rice distribution ratio, a classification threshold is automatically determined. This threshold is then incorporated into a preset extraction rule, and the multi-band SAR images from the current year are processed to generate a high-precision rice distribution map. This method not only considers the polarization value of a single phase (such as VV or VH), but also incorporates the backscatter coefficient changes of multiple phases. Furthermore, the backscatter coefficient combination characteristics of rice at different growth stages can be fully utilized, effectively improving the accuracy of rice identification and extraction, and achieving more accurate and reliable rice distribution mapping.

[0028] To this end, the present invention provides a rice extraction method based on multi-temporal SAR data, which can improve the accuracy of rice identification and extraction. The specific execution steps are as follows: Figure 1 As shown, including:

[0029] 101. Obtain multi-temporal SAR data corresponding to the target year and backscatter coefficients corresponding to the multi-temporal SAR data for the area to be extracted.

[0030] Among them, the target years include the current year and historical years.

[0031] It should be noted that in this embodiment, the region to be extracted can be one or more geographical areas where rice distribution monitoring is required. The current year refers to the year in which rice extraction is to be performed in the region to be extracted, while the historical year can be the year most recent to the current year for which complete multi-temporal historical SAR data is available for the region to be extracted.

[0032] In this step, SAR data covering the area to be extracted can be obtained from the satellite database. Common SAR data sources include the European Space Agency's Sentinel-1 and the Canadian Space Agency's RADARSAT series. Ensure that the selected data has sufficient temporal and spatial resolution to meet the needs of rice identification, for example, a spatial resolution of about 10 meters and a time interval ranging from one week to one month. Extract the backscatter coefficient of each phase SAR image, which usually includes values ​​under different polarization modes such as horizontal polarization (HH) and vertical polarization (VV). Arrange the SAR data in time series to form a multi-phase SAR dataset covering the entire growth cycle, including data from multiple key growth stages such as the transplanting period, jointing period, and filling period.

[0033] Specifically, the multi-temporal phases corresponding to rice can be determined by analyzing the phenological characteristics of the terrain in the region to be extracted. These phases include the transplanting, jointing, and filling periods. Based on these phases, first multi-temporal SAR data for the current year and second multi-temporal SAR data for historical years are specifically acquired. The first and second multi-temporal SAR data are then preprocessed to extract the corresponding backscatter coefficients. This preprocessing includes calibration, multi-look, filtering, polarization decomposition, and geocoding.

[0034] 102. According to the preset backscatter coefficient combination method of different time phases, the multi-phase SAR data are synthesized into a multi-band SAR image.

[0035] The multi-band SAR images include the first multi-band SAR images of the current year and the second multi-band SAR images of the historical years.

[0036] In this step, a backscatter coefficient combination for each time phase can be pre-set based on the physical structural changes of rice at different growth stages. This backscatter coefficient combination includes a sum, a ratio, and a radar vegetation index (Dprvivv). The sum is VV + VH, the ratio is VH / VV, and Dprvivv is 4 * VH / (VH + VV). Specifically, a table comparing different time phases and different backscatter coefficient combinations can be constructed and maintained to facilitate subsequent table lookup and determination.

[0037] The different polarization values ​​in the multi-temporal SAR data are combined according to the backscatter coefficient combination method to generate two sets of multi-band SAR images: one set is the first multi-band SAR image of the current year, and the other set is the second multi-band SAR image of the historical year. It should be noted that each band represents the backscattering characteristics of rice plants in a specific temporal phase. For example, the sum of the backscattering coefficients of the first transplanting period (VV + VH), the sum of the backscattering coefficients of the second transplanting period (VV + VH), the ratio of the backscattering coefficients of the first filling period (VH / VV), the backscattering coefficient of the jointing period (Dprvivv) (4*VH / (VH + VV)), and the backscattering coefficient of the second filling period (Dprvivv) (4*VH / (VH + VV)). By utilizing the combined features, the resulting multi-band SAR images can more comprehensively reflect the scattering characteristics of rice plants in different years and growth stages, capturing more details and features, which facilitates more accurate identification and extraction of rice plants.

[0038] 103. Based on the backscatter coefficient combination values ​​corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution, the backscatter coefficient combination classification threshold corresponding to each band is determined.

[0039] In this step, the backscatter coefficient combination values ​​for each band in the second multi-band SAR image from the historical year are analyzed, and statistical parameters such as the mean and standard deviation are calculated. The proportion of rice distribution is determined using historical rice distribution information (rice distribution maps or ground-based data, etc.) corresponding to the known historical years. The backscatter coefficient combination values ​​and rice distribution proportions corresponding to each band are then used to determine the backscatter coefficient combination classification threshold for each band. Specifically, a histogram can be drawn with the backscatter coefficient combination value corresponding to each band and the vertical axis as the proportion of the distribution number. Referring to the proportion of rice distribution corresponding to each band, troughs with similar proportions are selected from the troughs on both sides of each peak in the histogram as classification troughs, and the backscatter coefficient combination values ​​corresponding to these classification troughs are used as classification thresholds for each band. Alternatively, known ground-measured data, historical high-precision remote sensing data, and other rice distribution data can be used as a training set to train a random forest, support vector machine, neural network, etc. as a classifier. These classifiers can automatically learn the relationship between the backscatter coefficients of different bands and the rice distribution, and generate a classification threshold. This embodiment does not limit this.

[0040] In addition, in the process of determining the classification threshold by analyzing the histogram, factors such as the growth stage of rice and plant density can also be combined to make the backscatter coefficient combination classification threshold more accurate.

[0041] 104. Introduce the combined classification threshold of the backscatter coefficient corresponding to each band into the preset extraction rule, and use the extraction rule to extract rice from the first multi-band SAR image of the current year to obtain the rice distribution map corresponding to the extraction area in the current year.

[0042] In this step, according to the determined classification threshold, a specific extraction rule is formulated. The extraction rule can be set for each band, that is, one band corresponds to one extraction rule, so as to make each band more accurate when classifying and extracting rice. For example, it can be set that when the backscatter coefficient combination value of a certain pixel is greater than or less than a certain specific classification threshold, it is classified as rice. The formulated extraction rule is applied to the first multi-band SAR image of the current year, and it is judged pixel by pixel whether it belongs to the rice category. This process can use machine learning algorithms (such as random forests, support vector machines, etc.) to assist classification and further improve accuracy. The classification results of all pixels are combined to generate a final rice distribution map to show the specific distribution of rice in the area to be extracted.

[0043] It should be noted that in order to reduce background noise interference and ensure the accuracy of rice extraction, the first multi-band SAR image of the current year can also be initially segmented and denoised to remove background noise (such as non-agricultural land, water, etc.), so that the subsequent classification process can be more focused on rice, thereby improving the accuracy of classification and avoiding misclassification caused by noise. On this basis, since the segmented sub-images usually contain fewer types of ground objects and more uniform texture information, which makes features such as the backscattering coefficient more obvious, the new first multi-band SAR image obtained after the initial segmentation and denoising can be segmented again to obtain multiple new first multi-band SAR sub-images, and the above-mentioned extraction rules can be used to extract rice from the multiple new first multi-band SAR sub-images in sequence, which is conducive to distinguishing different types of ground objects, thereby improving the accuracy of rice classification and extraction.

[0044] Based on the above Figure 1 It can be seen from the implementation method that the present application provides a rice extraction method based on multi-temporal SAR data. When it is necessary to extract the distribution of rice, the multi-temporal SAR data corresponding to the current year and the historical year are first obtained. Then, according to the preset backscatter coefficient combination method of different phases, the multi-temporal SAR data are synthesized into a multi-band SAR image to obtain the first multi-band SAR image of the current year and the second multi-band SAR image of the historical year. Then, according to the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution, the backscatter coefficient combination classification threshold corresponding to each band is determined. Finally, the backscatter coefficient combination classification threshold corresponding to each band is introduced into the preset extraction rule, and the extraction rule is used to extract rice from the first multi-band SAR image of the current year to obtain the rice distribution map corresponding to the area to be extracted in the current year. The technical solution provided in this application introduces multi-temporal SAR data of the current year and historical years, and synthesizes the multi-temporal SAR data according to the combination of backscatter coefficients of different phases, so that the obtained multi-band SAR image can more comprehensively reflect the scattering characteristics of rice in different years and different growth stages, so that it can capture more details and features, which helps to more accurately identify and extract rice. The multi-band SAR images of historical years are used to determine the rice extraction threshold corresponding to the multi-band SAR image of the current year. Past experience and rules are fully considered, providing a reliable reference standard for rice extraction in the current year, reducing uncertainty and error, and accurately reflecting the rice distribution in the area to be extracted in the current year, thereby effectively improving the accuracy of rice identification and extraction.

[0045] Furthermore, the preferred embodiment of the present application is in the above Figure 1Based on this, the rice extraction process based on multi-temporal SAR data is described in detail. The specific steps are as follows: Figure 2 As shown, including:

[0046] 201. Obtain multi-temporal SAR data corresponding to the target year and backscatter coefficients corresponding to the multi-temporal SAR data for the area to be extracted.

[0047] This step is combined with the description of step 101 in the above method, and the same content will not be repeated here. It should be noted that the specific execution process of obtaining the multi-phase SAR data corresponding to the target year of the area to be extracted and the backscatter coefficient corresponding to the multi-phase SAR data is as follows: determining the multi-phase corresponding to rice according to the phenological characteristics of the ground objects in the area to be extracted, the multi-phase including the transplanting period, the jointing period, and the filling period; obtaining the SAR data corresponding to the target year of the area to be extracted according to the multi-phase to obtain the multi-phase SAR data, the multi-phase SAR data including the first multi-phase SAR data of the current year and the second multi-phase SAR data of the historical year; preprocessing the first multi-phase SAR data and the second multi-phase SAR data respectively to obtain the backscatter coefficient corresponding to each of the first multi-phase SAR data and the second multi-phase SAR data.

[0048] In this step, historical meteorological data, soil type, topography and other information of the area to be extracted are collected and analyzed in advance to understand the growth cycle of local major crops (especially rice). Specifically, reference can be made to agricultural science literature or information provided by local agricultural departments to clarify the main growth stages of rice in the area and their time distribution. At the same time, the key periods of rice growth are determined, including the transplanting period, the jointing period and the filling period. These periods usually correspond to specific farming activities and physiological changes and are important time nodes for monitoring and identifying rice. For example, in southern China, the transplanting period of rice is generally from April to May each year, the jointing period is from June to July, and the filling period is from August to September.

[0049] The key growth period determined above is used as the multi-temporal phase corresponding to rice to ensure that each key period has corresponding image coverage. According to the multi-temporal phase, suitable SAR data sources are screened from the public satellite database to obtain SAR data corresponding to the target year of the extracted area, including the current year and historical years. The time point of the SAR data is as close as possible to the center of the key growth period. All downloaded SAR images are arranged in time series to form a multi-temporal SAR dataset covering multiple key periods such as the transplanting period, jointing period, and filling period, obtaining the first multi-temporal SAR data and the second multi-temporal SAR data. The first multi-temporal SAR data and the second multi-temporal SAR data are preprocessed separately, including calibration, multi-viewing, filtering, polarization decomposition, geocoding, etc., and the backscatter coefficient of each phase SAR image is extracted as the basic feature for subsequent classification and analysis. The backscatter coefficient reflects the reflection intensity of the surface to the radar wave and is an important indicator for distinguishing different types of land features.

[0050] 202. According to the preset backscatter coefficient combination method of different time phases, the multi-phase SAR data are synthesized into a multi-band SAR image.

[0051] This step is combined with the description of step 102 in the above method, and the same contents will not be repeated here.

[0052] 203. Based on the backscatter coefficient combination values ​​corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution, the backscatter coefficient combination classification threshold corresponding to each band is determined.

[0053] This step is combined with the description of step 103 in the above method, and the same content will not be repeated here. It should be noted that the specific execution process of determining the backscatter coefficient combination threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of the distribution number corresponding to the backscatter coefficient combination value is as follows: based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image and the proportion of the distribution number corresponding to the backscatter coefficient combination value, a histogram corresponding to each band is drawn, the abscissa of the histogram is the backscatter coefficient combination value, and the ordinate is the proportion of the distribution number; with reference to the proportion of the rice distribution corresponding to each band, the classification trough corresponding to each band is determined in the troughs on both sides of each peak in the histogram corresponding to each band using the proximity principle, and the backscatter coefficient combination value corresponding to the classification trough is used as the backscatter coefficient combination threshold corresponding to each band.

[0054] In this step, the horizontal axis is set to the backscatter coefficient combination value, which should cover all possible values. The vertical axis is set to the proportion of the distribution, indicating the frequency of each backscatter coefficient value. Use professional geographic information system (GIS) software or a plotting library (such as Matplotlib or Seaborn) in a programming language (such as Python or R) to draw the histogram. For each band, draw a corresponding histogram to ensure that each histogram clearly shows the distribution of the backscatter coefficient values.

[0055] Key points representing rice distribution proportions are marked on the histogram to facilitate quick location during subsequent analysis. Simultaneously, the histogram for each band is analyzed to identify the troughs (local minima) on either side of each peak (local maximum). Peaks typically represent concentrated backscatter coefficient values, while troughs indicate the dividing points between different classes. Based on the rice distribution proportion, the trough closest to the rice proportion is identified within the troughs on either side of each peak. This trough is the classification trough, and troughs located between two distinct peaks are selected as the classification trough. The troughs located between two distinct peaks are selected as the classification troughs. The combined backscatter coefficient values ​​corresponding to each band classification trough are recorded; these values ​​serve as the classification threshold for each band. This threshold is used to distinguish rice from other land features and is a crucial component of subsequent classification rules. If multiple candidate troughs exist, the optimal classification threshold can be selected through further analysis (e.g., considering the stability of adjacent troughs and consistency with other bands).

[0056] It is worth noting that in this step, the combined backscatter coefficient values ​​corresponding to the troughs on both sides of the peak are selected as the classification threshold, which can establish a clear boundary between rice and other land features and enhance the contrast, so as to generate a high-quality rice distribution map, especially when the difference between rice and the background or other land features is not obvious.

[0057] For example, since the second multi-band SAR image synthesized above contains the backscatter coefficient combination values ​​of five bands, these five bands are 1, 2, 3, 4, and 5, where 1 is the first band of the transplanting period, 2 is the second band of the transplanting period, 3 is the first band of the filling period, 4 is the jointing period, and 5 is the second band of the filling period. Histograms are formed for each of the five bands, and the trough values ​​M on both sides of each peak are obtained. The trough values ​​of the backscatter coefficient combination at different phases are recorded as M nΔTn(n is the number of bands, n = 1, 2, 3, 4, 5, 1 and 2 are the sum of the backscattering coefficients corresponding to the two transplanting periods (VV + VH), 3 is the ratio of the backscattering coefficients of the corresponding phase of the first filling period (VH / VV), 4 is the backscattering coefficient Dprvivv (4*VH / (VH+VV)) of the corresponding phase of the jointing period, and 5 is the backscattering coefficient Dprvivv (4*VH / (VH+VV)) of the corresponding phase of the second filling period).

[0058] The determination method of ΔT is as follows:

[0059] Let X be the trough number of the backscatter coefficient combination statistical result curve. Count the backscatter coefficient combination values ​​corresponding to the five bands of the second multi-band SAR image to form a histogram (the horizontal axis is the backscatter coefficient combination value, and the vertical axis is the distribution ratio). The proportion of rice distribution corresponding to the second multi-band SAR image is recorded as A, the proportion as B, and the total proportion of backscatter coefficients in a certain section as C. The specific expression is:

[0060]

[0061] For each band n, select the X value when the backscatter coefficient segment ratio C is closest to the obtained rice planting ratio information.

[0062] Based on the determined X value, the ΔT corresponding to each band is determined. The specific expression is:

[0063]

[0064] Substitute the ΔT corresponding to each band obtained into M nΔTn , and obtain the corresponding classification threshold.

[0065] 204. Perform initial segmentation on the first multi-band SAR image according to a first preset scale to obtain a plurality of first multi-band SAR sub-images.

[0066] In this step, the first preset scale can be specifically determined based on factors such as the spatial resolution of the image and the size of the object. This scale determines the size of each sub-image after the initial segmentation. For example, the first preset scale is 0.4 (shape), 0.5 (density), and the scale parameter is 1. Based on the characteristics of the SAR image, a suitable segmentation algorithm is selected, such as pixel-based fixed window segmentation or region growing-based segmentation. Using the selected segmentation algorithm, the first multi-band SAR image is segmented according to the first preset scale to obtain multiple non-overlapping first multi-band SAR sub-images.

[0067] 205. Calculate a combined backscatter coefficient denoising threshold based on minimum backscatter coefficients of various types of ground objects in the plurality of first multi-band SAR sub-images.

[0068] The backscatter coefficient combined denoising threshold is used to remove background noise in the plurality of first multi-band SAR sub-images.

[0069] It should be noted that the backscatter coefficient combination denoising threshold is calculated based on the minimum value of the backscatter coefficient combination values ​​of various ground objects in multiple first multi-band SAR sub-images. The specific expression is:

[0070] ΔB <N×T min ;

[0071] Where ΔB is the combined denoising threshold of backscatter coefficients used to remove background noise, T min It is the minimum value among the combined values ​​of backscatter coefficients of various types of ground objects, and N is a natural number greater than or equal to 2.

[0072] In this step, various types of objects include but are not limited to paddy fields, buildings, forests, etc. The choice of N depends on the intensity of background noise and the degree of detail of the objects that need to be retained. Traverse all sub-images and record the minimum value in each category to find the minimum value T of the backscatter coefficient of each type of object in all sub-images min , and substitute it into the above expression to get the combined denoising threshold of the backscatter coefficient. In general, ΔB usually needs to be much larger than T min The choice of N depends on the intensity of background noise and the degree of detail of the ground objects that need to be retained. Specifically, 8, 10, etc. can be selected to ensure that ΔB can effectively reduce noise without accidentally deleting important ground object information.

[0073] 206. Remove background noise from the plurality of first multi-band SAR sub-images according to the combined denoising threshold of the backscatter coefficient, and merge the plurality of denoised first multi-band SAR sub-images to obtain a new first multi-band SAR image of the current year.

[0074] In this step, the calculated combined backscatter coefficient denoising threshold ΔB is applied to each first multiband SAR sub-image. By comparing the backscatter coefficient of each pixel with ΔB, a decision is made as to whether to retain the pixel's information. Pixels below ΔB are treated as background noise and removed, while pixels above or equal to ΔB retain their information. This process can be achieved using binarization or other filtering techniques. All denoised first multiband SAR sub-images are reassembled into a complete image, ensuring seamless integration between components, to form the new first multiband SAR image for the current year.

[0075] 207 . Perform secondary segmentation on the new first multi-band SAR image according to a second preset scale to obtain a plurality of new first multi-band SAR sub-images.

[0076] The second preset scale is larger than the first preset scale.

[0077] In this step, the second preset scale can also be determined based on factors such as the image's spatial resolution and feature size. However, it is typically larger than the first preset scale because a larger segmentation scale can help capture feature variations or patterns over a wider range, thus accommodating different levels of analysis. The new first multi-band SAR image is then re-segmented at the second preset scale using the same segmentation algorithm used for the initial segmentation, generating multiple new first multi-band SAR sub-images. This facilitates execution of step 208.

[0078] 208. Introduce the backscatter coefficient combination threshold corresponding to each band into the preset extraction rule, and use the extraction rule to extract rice from multiple new first multi-band SAR sub-images to obtain the rice distribution map corresponding to the extraction area in the current year.

[0079] This step is combined with the description of step 207 in the above method. It is only necessary to replace the first multi-band SAR image with a plurality of new first multi-band SAR sub-images. Therefore, the same contents will not be repeated here.

[0080] It should be noted that the backscatter coefficient combination classification threshold corresponding to each band is introduced into the preset extraction rule, and the extraction rule is used to extract rice from the first multi-band SAR image of the current year to obtain the rice distribution map corresponding to the area to be extracted in the current year. The specific execution process is: the backscatter coefficient combination classification threshold corresponding to the transplanting period band and the jointing period band is used as the extraction upper limit, and the backscatter coefficient combination classification threshold corresponding to the filling period band is used as the extraction lower limit to obtain the rice extraction results corresponding to each band; based on the rice extraction results corresponding to each band, the rice distribution map corresponding to the area to be extracted in the current year is generated.

[0081] Among them, the various bands are the transplanting period band, the jointing period band and the filling period band.

[0082] In this step, the combined classification thresholds for the backscatter coefficients corresponding to the transplanting and jointing bands are used as the upper extraction limit. That is, if the backscatter coefficient of a pixel is below these two thresholds, it is not classified as rice. The combined classification threshold for the backscatter coefficients corresponding to the filling band is used as the lower extraction limit. That is, if the backscatter coefficient of a pixel is above this threshold, it is considered likely to be rice. Based on these upper and lower limits, specific extraction rules are formulated. For example, a logical expression can be defined to determine whether each pixel belongs to the rice category. Using these extraction rules, each pixel in the first multi-band SAR image of the current year is classified based on its backscatter coefficients in the transplanting, jointing, and filling bands. The classification results of all pixels are summarized to form a complete rice distribution map. During this process, spatial filtering or morphological operations (such as dilation and erosion) can be applied to improve the spatial coherence of the classification results and reduce the impact of isolated noise points. Smoothing algorithms (such as Gaussian filtering) can also be used to process the boundaries of the classification results to make the rice distribution map smoother and more natural.

[0083] Continuing with the example in step 203, for example, combined with the above description, since the transplanting period band includes the transplanting period band I and the transplanting period band II, and the filling period band includes the filling period band I and the filling period band II, the rice extraction is performed based on the multiple first multi-band SAR sub-images after secondary segmentation. The specific logical expression is as follows:

[0084] S t1 ≤M 1ΔT1 ; ①

[0085] S t2 ≤M 2ΔT2 ; ②

[0086] R t4 ≥M 3ΔT3 ; ③

[0087] D t3 ≤M 4ΔT4 ④

[0088] D t4 ≥M 5ΔT5 ⑤

[0089] ①∪②∩③∩④∩⑤

[0090] Where S is the sum of the backscattering coefficients of the corresponding phase (VV + VH), R is the ratio of the backscattering coefficients of the corresponding phase (VH / VV), D is the backscattering coefficient Dprvivv (4*VH / (VH+VV)) of the corresponding phase, t1 and t2 are the phases corresponding to the two transplanting stages, t3 is the phase corresponding to the jointing stage, and t4 and t5 are the phases corresponding to the two filling stages.

[0091] Furthermore, in order to further remove interference noise in the rice distribution map and ensure the accuracy of the rice distribution map, specifically, after introducing the backscatter coefficient combination classification threshold corresponding to each band into a preset extraction rule, and using the extraction rule to extract rice from the first multi-band SAR image of the current year, and obtaining the rice distribution map corresponding to the current year of the area to be extracted, the method also includes: calculating the number of pixels occupied by each rice patch in the rice distribution map; determining a pixel number threshold for removing interference noise based on the resolution corresponding to the first multi-temporal SAR data and a preset result accuracy requirement, the pixel number threshold being used to remove interference noise in the rice patches; defining rice patches with a pixel number less than or equal to the pixel number threshold as noise patches, and marking the pixels occupied by the noise patches as non-rice pixels, to obtain the final rice distribution map.

[0092] In this step, connected component analysis or region growing algorithms can be used to identify all independent rice patches in the rice distribution map. For each patch, the number of pixels contained in it is counted and this information is recorded. Specifically, a table or database can be created to store the unique identifier, location information (such as center coordinates), number of pixels, and other attributes of each patch. The resolution corresponding to the first multi-phase SAR data is analyzed to understand the actual ground area represented by each pixel. Based on the preset result accuracy requirements, an acceptable maximum error range is defined to ensure that the final result meets the application requirements. Based on the above analysis results, a reasonable pixel number threshold is set as the standard for distinguishing real rice patches from noise patches. The number of pixels in each patch is compared with the set pixel number threshold. Patches with a pixel number less than or equal to the pixel number threshold are defined as noise patches. The pixels occupied by these noise patches can be re-marked as non-rice pixels on the original rice distribution map. At the same time, the corresponding records in the patch attribute table are modified to remove the patch information marked as noise. According to the updated patch attribute table, the rice distribution map is reconstructed to ensure that only real rice patches with a pixel number greater than the threshold are retained to obtain the final rice distribution map.

[0093] It should be noted that the pixel number threshold for removing interference noise is determined according to the resolution corresponding to the first multi-temporal SAR data and the preset result accuracy requirement. The specific expression is:

[0094]

[0095] Where P is the pixel number threshold for removing interference noise, Q is the result accuracy requirement, specifically referring to the minimum ground area that each rice patch should cover in the desired result (for example, in square meters), and E is the resolution corresponding to the first multi-temporal SAR data, specifically referring to the actual ground length represented by each pixel (for example, in meters / pixel).

[0096] Through the specific expression given in this step, the interference noise can be further effectively removed from the rice distribution map, the quality of the final classification result can be improved, and the accuracy of rice distribution can be ensured.

[0097] Furthermore, as a response to the above Figure 1-2 The embodiment of the method shown in the figure is implemented. The embodiment of the present application provides a rice extraction device based on multi-temporal SAR data, which is used to improve the accuracy of rice identification and extraction. The embodiment of the device corresponds to the aforementioned method embodiment. For ease of reading, this embodiment will no longer repeat the details of the aforementioned method embodiment one by one, but it should be clear that the device in this embodiment can correspond to all the contents of the aforementioned method embodiment. Specifically, Figure 3 As shown, the device includes:

[0098] An acquisition unit 301 is configured to acquire multi-temporal SAR data corresponding to a target year of an area to be extracted and a backscatter coefficient corresponding to the multi-temporal SAR data, wherein the target year includes a current year and a historical year;

[0099] a processing unit 302 configured to synthesize the multi-temporal SAR data obtained by the acquisition unit 301 into a multi-band SAR image according to a combination of backscatter coefficients preset for different temporal phases, wherein the multi-band SAR image includes a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year;

[0100] A first determining unit 303 is configured to determine a backscatter coefficient combination classification threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year obtained by the processing unit 302 and the proportion of rice distribution;

[0101] The extraction unit 34 is configured to introduce the backscatter coefficient combination classification threshold corresponding to each band obtained by the first determination unit 303 into a preset extraction rule, and use the extraction rule to extract rice from the first multi-band SAR image of the current year obtained by the processing unit, thereby obtaining a rice distribution map corresponding to the area to be extracted in the current year.

[0102] Further, such as Figure 4 As shown, the acquisition unit 301 includes:

[0103] The first determining module 3011 is configured to determine multiple temporal phases corresponding to rice according to the phenological characteristics of the landforms in the area to be extracted, wherein the multiple temporal phases include the transplanting period, the jointing period, and the filling period;

[0104] An acquisition module 3012 is configured to acquire SAR data corresponding to the target year for the area to be extracted according to the multi-temporal phases obtained by the first determination module 3011, to obtain the multi-temporal SAR data, wherein the multi-temporal SAR data includes first multi-temporal SAR data for the current year and second multi-temporal SAR data for the historical year;

[0105] The preprocessing module 3013 is configured to preprocess the first multi-temporal SAR data and the second multi-temporal SAR data obtained by the acquisition module 3012 to obtain backscatter coefficients corresponding to the first multi-temporal SAR data and the second multi-temporal SAR data, respectively.

[0106] Further, such as Figure 4 As shown, the first determining unit 303 includes:

[0107] a drawing module 3031 for drawing a histogram corresponding to each band in the second multi-band SAR image based on the backscatter coefficient combination values ​​corresponding to each band and the proportion of the distribution quantity corresponding to the backscatter coefficient combination values, wherein the abscissa of the histogram is the backscatter coefficient combination value and the ordinate is the proportion of the distribution quantity;

[0108] The second determination module 3032 is used to refer to the proportion of rice distribution corresponding to each band, and use the proximity principle to determine the classification troughs corresponding to each band in the troughs on both sides of each peak in the histogram corresponding to each band obtained by the drawing module 3031, and use the backscatter coefficient combination value corresponding to the classification trough as the backscatter coefficient combination threshold corresponding to each band.

[0109] Further, such as Figure 4 As shown, the device also includes:

[0110] A first segmentation unit 305 is configured to perform initial segmentation on the first multi-band SAR image according to a first preset scale before the extraction unit 304 to obtain a plurality of first multi-band SAR sub-images;

[0111] A first calculation unit 306 is configured to calculate a backscattering coefficient combination denoising threshold based on minimum backscattering coefficients of various types of ground objects in the plurality of first multi-band SAR sub-images obtained by the first segmentation unit 305, wherein the backscattering coefficient combination denoising threshold is used to remove background noise in the plurality of first multi-band SAR sub-images;

[0112] a first denoising unit 307 configured to remove background noise from each of the plurality of first multi-band SAR sub-images according to the backscatter coefficient combined denoising threshold value obtained by the first calculating unit 306, and merge the plurality of denoised first multi-band SAR sub-images to obtain a new first multi-band SAR image of the current year;

[0113] a second segmentation unit 308 configured to perform secondary segmentation on the new first multi-band SAR image obtained by the first denoising unit 307 according to a second preset scale to obtain a plurality of new first multi-band SAR sub-images, wherein the second preset scale is larger than the first preset scale;

[0114] The extraction unit 304 is specifically configured to:

[0115] The backscatter coefficient combination threshold corresponding to each band is introduced into a preset extraction rule, and the extraction rule is used to extract rice from the multiple new first multi-band SAR sub-images obtained by the second segmentation unit 308 to obtain a rice distribution map corresponding to the area to be extracted in the current year.

[0116] Further, such as Figure 4 As shown, the first calculation unit 306 is specifically expressed as follows:

[0117] ΔB <N×T min ;

[0118] Where ΔB is the combined denoising threshold of backscatter coefficients used to remove background noise, T min It is the minimum value among the combined values ​​of backscatter coefficients of various types of ground objects, and N is a natural number greater than or equal to 2.

[0119] Further, such as Figure 4 As shown, the wavebands are respectively a waveband of the transplanting period, a waveband of the jointing period, and a waveband of the filling period; the extraction unit 304 includes:

[0120] An extraction module 3041 is configured to use the backscatter coefficient combination classification threshold corresponding to the transplanting period band and the jointing period band as an extraction upper limit, and use the backscatter coefficient combination classification threshold corresponding to the filling period band as an extraction lower limit, to obtain rice extraction results corresponding to each band;

[0121] The generating module 3042 is configured to generate a rice distribution map corresponding to the area to be extracted in the current year based on the rice extraction results corresponding to the various bands obtained by the extracting module 3041 .

[0122] Further, such as Figure 4 As shown, the device also includes:

[0123] A second calculation unit 309 is used to calculate the number of pixels occupied by each rice patch in the rice distribution map after the extraction unit 304;

[0124] A second determining unit 310 is configured to determine a pixel number threshold for removing interference noise according to a resolution corresponding to the first multi-temporal SAR data and a preset result accuracy requirement, wherein the pixel number threshold is used for removing interference noise in the rice patch;

[0125] The second denoising unit 311 is used to define the rice patches whose number of pixels obtained by the second calculating unit 309 is less than or equal to the pixel number threshold obtained by the second determining unit 310 as noise patches, and mark the pixels occupied by the noise patches as non-rice pixels, to obtain a final rice distribution map.

[0126] Further, such as Figure 4 As shown, the second determining unit 310 is specifically expressed as follows:

[0127]

[0128] Where P is the pixel number threshold for removing interference noise, Q is the result accuracy requirement, and E is the resolution corresponding to the first multi-temporal SAR data.

[0129] Furthermore, the embodiment of the present application also provides a storage medium, which is used to store a computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the above Figure 1-2 The rice extraction method based on multi-temporal SAR data described in.

[0130] Furthermore, the embodiment of the present application also provides a processor, which is used to run a program, wherein the program executes the above Figure 1-2 The rice extraction method based on multi-temporal SAR data described in.

[0131] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0132] It is understood that the relevant features of the above methods and devices can be referenced to each other. In addition, the terms "first" and "second" in the above embodiments are used to distinguish between the embodiments, and do not represent the advantages and disadvantages of the embodiments.

[0133] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0134] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems may also be used together with the teachings herein. Based on the above description, it is apparent that the structure required for constructing such systems is suitable. In addition, the present application is not directed to any specific programming language. It should be understood that various programming languages ​​may be utilized to implement the present application described herein, and the description of the specific languages ​​above is provided for the purpose of disclosing the preferred embodiment of the present application.

[0135] In addition, the memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0136] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0137] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0138] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0139] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0140] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0141] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0142] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0143] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0144] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A rice extraction method based on multi-temporal SAR data, characterized in that: The method comprises: Acquire multi-temporal SAR data corresponding to a target year for the area to be extracted and a backscatter coefficient corresponding to the multi-temporal SAR data, wherein the target year includes a current year and a historical year; synthesizing the multi-temporal SAR data into a multi-band SAR image according to a combination of backscatter coefficients preset for different temporal phases, wherein the multi-band SAR image includes a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year; Determining a backscatter coefficient combination classification threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution; Introducing the backscatter coefficient combination classification threshold corresponding to each band into a preset extraction rule, and using the extraction rule to extract rice from the first multi-band SAR image of the current year to obtain a rice distribution map corresponding to the area to be extracted in the current year; Determine the backscatter coefficient combination threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year and the proportion of rice distribution, including: Based on the backscatter coefficient combination values ​​corresponding to each band in the second multi-band SAR image and the proportion of the distribution quantity corresponding to the backscatter coefficient combination values, a histogram corresponding to each band is drawn, where the abscissa of the histogram is the backscatter coefficient combination value, and the ordinate is the proportion of the distribution quantity; Referring to the proportion of rice distribution corresponding to each band, the classification trough corresponding to each band is determined in the troughs on both sides of each peak in the histogram corresponding to each band using the proximity principle, and the backscatter coefficient combination value corresponding to the classification trough is used as the backscatter coefficient combination threshold corresponding to each band.

2. The method according to claim 1, characterized in that Acquiring multi-temporal SAR data corresponding to the target year for the area to be extracted and backscatter coefficients corresponding to the multi-temporal SAR data, including: Determining the multi-temporal phases corresponding to rice according to the phenological characteristics of the landforms in the area to be extracted; Acquire SAR data corresponding to the target year for the area to be extracted according to the multi-temporal phases to obtain the multi-temporal SAR data, wherein the multi-temporal SAR data includes first multi-temporal SAR data of the current year and second multi-temporal SAR data of the historical year; Preprocessing is performed on the first multi-temporal SAR data and the second multi-temporal SAR data respectively to obtain backscatter coefficients corresponding to the first multi-temporal SAR data and the second multi-temporal SAR data.

3. The method according to claim 1, characterized in that Before introducing the backscatter coefficient combination classification threshold corresponding to each band into a preset extraction rule and performing rice extraction on the first multi-band SAR image of the current year using the extraction rule to obtain a rice distribution map corresponding to the current year for the area to be extracted, the method further includes: performing an initial segmentation on the first multi-band SAR image according to a first preset scale to obtain a plurality of first multi-band SAR sub-images; Calculating a backscattering coefficient combination denoising threshold based on minimum backscattering coefficients of various types of ground objects in the plurality of first multi-band SAR sub-images, wherein the backscattering coefficient combination denoising threshold is used to remove background noise in the plurality of first multi-band SAR sub-images; removing background noise from a plurality of first multi-band SAR sub-images according to the backscatter coefficient combined denoising threshold, and merging the plurality of denoised first multi-band SAR sub-images to obtain a new first multi-band SAR image of the current year; performing secondary segmentation on the new first multi-band SAR image according to a second preset scale to obtain a plurality of new first multi-band SAR sub-images, wherein the second preset scale is larger than the first preset scale; The step of introducing the backscatter coefficient combination threshold corresponding to each band into a preset extraction rule, and using the extraction rule to extract rice from the first multi-band SAR image of the current year to obtain a rice distribution map corresponding to the current year for the area to be extracted comprises: The backscatter coefficient combination threshold corresponding to each band is introduced into a preset extraction rule, and the extraction rule is used to extract rice from the multiple new first multi-band SAR sub-images to obtain a rice distribution map corresponding to the area to be extracted in the current year.

4. The method according to claim 3, characterized in that The backscatter coefficient combination denoising threshold is calculated based on the minimum value of the backscatter coefficient combination values ​​of various ground objects in the plurality of the first multi-band SAR sub-images. The specific expression is: ΔB <N×T min ; Where ΔB is the combined denoising threshold of backscatter coefficients used to remove background noise, T min It is the minimum value among the combined values ​​of backscatter coefficients of various types of ground objects, and N is a natural number greater than or equal to 2.

5. The method according to any one of claims 1 to 4, characterized in that The bands are a transplanting period band, a jointing period band, and a filling period band; the backscatter coefficient combination classification threshold corresponding to each band is introduced into a preset extraction rule, and the extraction rule is used to extract rice from the first multi-band SAR image of the current year, thereby obtaining a rice distribution map corresponding to the current year for the area to be extracted, including: The backscatter coefficient combination classification threshold corresponding to the transplanting period band and the jointing period band is used as the extraction upper limit, and the backscatter coefficient combination classification threshold corresponding to the filling period band is used as the extraction lower limit, to obtain the rice extraction results corresponding to each band; A rice distribution map corresponding to the area to be extracted in the current year is generated based on the rice extraction results corresponding to each band.

6. The method according to any one of claims 1 to 4, characterized in that After introducing the backscatter coefficient combination classification threshold corresponding to each band into a preset extraction rule and using the extraction rule to extract rice from the first multi-band SAR image of the current year to obtain a rice distribution map corresponding to the current year in the area to be extracted, the method further includes: Calculating the number of pixels occupied by each rice patch in the rice distribution map; determining a pixel number threshold for removing interference noise according to a resolution corresponding to the first multi-temporal SAR data and a preset result accuracy requirement, wherein the pixel number threshold is used for removing interference noise in the rice patch; The rice patches whose pixel number is less than or equal to the pixel number threshold are defined as noise patches, and the pixels occupied by the noise patches are marked as non-rice pixels, to obtain a final rice distribution map.

7. A rice extraction device based on multi-temporal SAR data, characterized in that: The device comprises: An acquisition unit, configured to acquire multi-temporal SAR data corresponding to a target year of an area to be extracted and a backscatter coefficient corresponding to the multi-temporal SAR data, wherein the target year includes a current year and a historical year; a processing unit configured to synthesize the multi-temporal SAR data obtained by the acquisition unit into a multi-band SAR image according to a combination of backscatter coefficients preset for different temporal phases, wherein the multi-band SAR image includes a first multi-band SAR image of the current year and a second multi-band SAR image of the historical year; a first determining unit, configured to determine a backscatter coefficient combination classification threshold corresponding to each band based on the backscatter coefficient combination value corresponding to each band in the second multi-band SAR image of the historical year obtained by the processing unit and the proportion of rice distribution; an extraction unit, configured to introduce the backscatter coefficient combination classification threshold corresponding to each band obtained by the first determination unit into a preset extraction rule, and perform rice extraction on the first multi-band SAR image of the current year obtained by the processing unit using the extraction rule, to obtain a rice distribution map corresponding to the current year for the area to be extracted; The first determining unit includes: a drawing module, configured to draw a histogram corresponding to each band based on the backscatter coefficient combination values ​​corresponding to each band in the second multi-band SAR image and the proportion of the distribution quantity corresponding to the backscatter coefficient combination values, wherein the abscissa of the histogram is the backscatter coefficient combination value and the ordinate is the proportion of the distribution quantity; The second determination module is used to refer to the proportion of rice distribution corresponding to each band, and use the proximity principle to determine the classification troughs corresponding to each band in the troughs on both sides of each peak in the histogram corresponding to each band obtained by the drawing module, and use the backscatter coefficient combination value corresponding to the classification trough as the backscatter coefficient combination threshold corresponding to each band.

8. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is executed, the device where the storage medium is located is controlled to execute the rice extraction method based on multi-temporal SAR data according to any one of claims 1 to 6.

9. A processor, characterized in that: The processor is configured to run a program, wherein the program, when running, executes the rice extraction method based on multi-temporal SAR data according to any one of claims 1 to 6.

Images