Rice and shrimp field identification method based on seasonal facies characteristics

Abstract

The invention discloses a rice and shrimp field identification method based on seasonal facies characteristics, and relates to the field of farmland land utilization remote sensing image classification, and the method comprises the steps: rice field radar time sequence data acquisition, radar time sequence data synthesis, missing value interpolation and filtering smoothing, rice field dynamic key point time extraction, and time sequence characteristic parameter extraction. And performing comparison and judgment based on the key points and the characteristic parameters. Therefore, the problem that optical image data acquisition in cloudy and foggy areas is limited is solved, and automatic rice and shrimp field identification based on radar data is effectively realized.

Description

Technical Field

[0001] The invention relates to the field of farmland land utilization remote sensing image classification, and in particular to a rice-shrimp field identification method based on seasonal characteristics. Background Art

[0002] Rice-shrimp farming is an ecological farming model that organically combines rice planting with crayfish farming, using the three-dimensional space of rice fields to plant one crop of rice throughout the year and raise crayfish. Its state changes over time during the year, showing the characteristics of water area-rice-water area. Compared with conventional rice fields, the synergistic effect of the rice-shrimp symbiotic system can improve the ecological environment of rice fields and significantly reduce CH4 emissions. At the same time, it can also achieve stable rice production and increase income from aquatic products, with significant social, economic and ecological benefits. The distribution and area of ​​rice-shrimp fields are obtained based on remote sensing time series data, providing a reference for regional rice yield estimation and rice field CH4 emission assessment, which meets the needs of low-carbon agricultural green development with data as the key production factor.

[0003] Existing research on remote sensing identification of rice-shrimp fields mainly uses medium-to-high resolution optical remote sensing data. The weather in the southwest region is often cloudy and foggy, making it difficult to obtain high-quality optical images. The planting structure in the area is complex, and it is difficult to distinguish it from lotus fields with similar seasonal characteristics using only vegetation and flooding signals, which limits the remote sensing monitoring capabilities and accuracy of rice-shrimp fields in the region. Summary of the invention

[0004] The purpose of the present invention is to provide a rice-shrimp field identification method based on time series radar remote sensing data, so as to overcome the problem of difficulty in obtaining optical data in foggy areas and improve the identification accuracy and adaptability of rice-shrimp fields.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: Acquisition of time series data of paddy field radar; Radar time series data synthesis, missing value interpolation and filtering smoothing; Extract the key points of paddy field dynamics; Extraction of time series feature parameters; Compare and judge based on key points and feature parameters.

[0006] In some embodiments, the paddy field radar time series data acquisition includes acquiring all radar time series data throughout the year and the radar time series data of the paddy field, wherein the radar time series data is the ratio of the intensity of the radar signal to the intensity of the incident signal, that is, the backscatter coefficient, which is usually expressed in decibels (dB). The paddy field radar time series data is the average of the radar signals of all pixels in each period within the plot.

[0007] In some embodiments, the rice field plot is the smallest basic unit for growing rice.

[0008] In some embodiments, the radar time series data includes backscatter coefficients, which are used to detect the dynamic changes of rice fields from water bodies to rice and then back to water bodies within a year.

[0009] In some embodiments, the radar time series data synthesis, missing value interpolation and filtering smoothing include: equally spaced synthesis of radar backscatter time series data to generate backscatter time series data with the same time interval, and then filling the missing values ​​through linear interpolation, and smoothing the radar backscatter signal time series data through the SG filtering algorithm.

[0010] In some embodiments, the paddy field dynamic key point time includes the paddy field state and seasonal phase dynamic change key points, and the paddy field state includes: paddy field state S1 before rice transplanting, rice growth state S2, and paddy field state S3 after rice harvesting. The extraction of paddy field dynamic key point time includes the time t1 of rice transplanting and the time t2 of rice harvesting, which is extracted according to the slope change characteristics of the paddy field radar backscattering coefficient curve.

[0011] In some embodiments, the paddy field states S1, S2, and S3 are represented by continuous integer digital codes, and their time series data use radar data to extract backscatter coefficients.

[0012] In some embodiments, the slope of the rice field radar backscatter coefficient curve is calculated using a difference method.

[0013] In some embodiments, the time series characteristic parameters include: the mean and variance of the backscatter coefficient of the rice field within the year and the mean of the S1 state, which are calculated and obtained based on the backscatter coefficient curve of the rice field.

[0014] In some embodiments, the annual backscatter coefficient mean and variance of the rice field, including statistical feature analysis of radar time series data, are used to evaluate the concentration and dispersion of the data to assist in crop classification, water body monitoring and its changing process.

[0015] In some embodiments, based on the comparison of key points and characteristic parameters, including the time when t1 and t2 occur, and the comparison of characteristic parameters such as the range of change, mean, and variance of the radar subsequent scattering coefficient of the rice field, it is determined to be a rice-shrimp field if the conditions are met.

[0016] The beneficial effects that may be brought about by a rice-shrimp field identification method based on seasonal characteristics disclosed in this application include but are not limited to: The final result is the radar signal change process and duration that dynamically reflects the seasonal changes in the rice fields. The rice-shrimp fields are identified based on the duration of the rice fields, the range of seasonal radar feature changes, and the radar signals in the full shrimp farming state. The data processed by the present invention quantitatively describes the radar signal change process of the rice-shrimp fields in different seasonal phases, reduces radar signal noise, obtains more reasonable radar time series data of the rice-shrimp fields, and realizes remote sensing identification of the rice-shrimp fields. Compared with optical image data, the present invention overcomes the problem of limited acquisition of optical image data in cloudy and foggy areas, and improves the remote sensing monitoring capability of rice-shrimp fields in cloudy and foggy areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a rice-shrimp field identification method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the key change recognition principle of the time series data characteristics of the rice-shrimp field radar according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the time series data of the original radar backscattering of a rice-shrimp field according to an embodiment of the present invention; Figure 4 It is a schematic diagram of radar backscattering time series data after synthetic interpolation of rice-shrimp fields according to an embodiment of the present invention; Figure 5 A schematic diagram of characteristic values ​​of backscatter coefficients of rice-shrimp fields obtained through data synthesis, interpolation and smoothing according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the time series data of the rice-shrimp field status obtained through data processing according to an embodiment of the present invention; Figure 7 Schematic diagram of a method for calculating the time of key points of changes in state characteristics of a rice-shrimp field according to an embodiment of the present invention; Figure 8 Schematic diagram of a method for identifying rice-shrimp fields according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present application more clear, the present application is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0019] On the contrary, the present application covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present application as defined by the claims. Further, in order to make the public have a better understanding of the present application, some specific details are described in detail in the detailed description of the present application below. Those skilled in the art can fully understand the present application without the description of these details.

[0020] The following is a detailed description of a rice-shrimp field identification method based on seasonal characteristics involved in the embodiments of the present application. It should be noted that the following embodiments are only used to explain the present application and do not constitute a limitation of the present application.

[0021] like Figure 1-8 As shown, a rice-shrimp field identification method based on seasonal characteristics comprises the following steps: Step 1: Obtaining radar time series data of paddy fields. By obtaining radar time series data of the area to be identified throughout the year, masking is performed based on paddy field data to extract all pixels in the plot. The radar time series signal corresponding to time t And find the mean , thereby obtaining the corresponding time series of the plot Radar time series data .

[0022] Step 2: radar time series data synthesis, missing value interpolation and filtering smoothing; according to The range of values ​​generates equally spaced time series ,Will Synthesize equally spaced time series, perform linear interpolation on missing values, and then smooth the time series data using the SG filtering algorithm to obtain a time series curve ; Step 3, extracting the key point time of the paddy field dynamics, wherein the key point of the paddy field dynamics includes the paddy field state and the key point time.

[0023] The paddy field status includes: paddy field status S1 before rice transplanting, rice growth status S2, and paddy field status S3 after rice harvesting; The key time points include: rice transplanting time t1, rice harvesting time t2; Determine t1 and t2 according to the change of the rice field time series curve. The start time of the rice growth state S2 corresponds to t1, and the start time of the rice field state S3 after rice harvest corresponds to t2, thereby determining S1, S2, and S3; Determine t1 and t2 according to the slope of the rice field radar time series curve. The slope at time t in the time series is given by the formula It is calculated that when the slope increases rapidly from approximately 0.5, it corresponds to t1, and when the slope changes from a positive value to a negative value and decreases rapidly, it corresponds to t2.

[0024] Step 4: Extract time series feature parameters; Based on the rice field radar time series data Calculate the mean and variance , the mean S1 of state S1 mean , in order to quantitatively describe the dynamic change characteristics of radar time series data of rice fields in different seasonal phases.

[0025] Step 5: Compare and judge based on key points and feature parameters. 1、 t2 and characteristic parameters as the benchmark, by comparing with the key points and characteristic parameters of normal rice-shrimp fields, when the transplanting period is t 1、 The maturity period t2 is around 135 and 265, with a deviation of ±10 days. It was judged to be a rice-shrimp field.

[0026] The following is an explanation using a specific embodiment as an example:

[0027] Embodiment 1: The time series SAR remote sensing images of the VH polarization data of the Sentinel-1 GRD product in the year of the area to be identified are obtained, and the images are preprocessed, including boundary noise removal, speckle filtering, optimized terrain correction, backscatter coefficient conversion, etc. In this embodiment, the dynamic change characteristics of the rice field are represented by the time series data of the backscatter coefficient of the rice field radar.

[0028] Step 1, obtaining the radar time series data of the paddy field. On the basis of obtaining the radar time series data of the area to be identified throughout the year, the deep learning field segmentation model obtained by training the high-resolution remote sensing image of the labeled cultivated field sample data set is used to extract the field boundary range of the area to be identified, and then the radar time series data of all pixels in each field are extracted one by one, and the average value of the radar backscatter coefficient at each moment is calculated to obtain the backscatter coefficient time series data of each field. The radar time series data of the paddy field is screened out according to the condition that the minimum value of the time series data should be less than -20; Step 2: Synthesize the 10-day mean of the time series data of each rice field, perform linear interpolation on the missing values ​​at a certain moment, use the SG filtering algorithm to smooth the time series data, further reduce the noise on the time series data, and obtain the rice field backscattering time series curve. The effect is shown in Figure 2 ; Step 3: Extract the dynamic key points of the paddy field according to Figure 2 The principle of identifying key changes in seasonal characteristics of rice-shrimp fields is to determine t1 and t2 from the changes in the time series data of the rice growth process. The slope change point of the curve calculated based on the backscattering time series data of the rice field is used to determine t1 and t2. The specific time point extraction is based on the following: the minimum time when Slope ≥ 0.5 and the next moment Slope is greater than or equal to 1 is t1, and the minimum time when Slope ≤ 0 and the next moment Slope ≤ -1 is t2. For the schematic diagram of determining the dynamic key point time of the rice field, see Figure 4 , determine states S1, S2, S3 according to t1 and t2, where t1 is the maximum moment of state S1 and the minimum moment of S2, and t2 is the maximum moment of state S2 and the minimum moment of S3; Step 4: Extract time series characteristic parameters and calculate the mean backscatter coefficient based on the radar time series data of the field. and variance , the mean S1 of state S1 mean .

[0029] Step 5: Compare and judge based on key points and feature parameters, see Figure 8 , the growth time Δt is calculated based on t2 and t1, and the growth time Δt and S1 of the normal rice-shrimp field are compared. mean For comparison, when growing in the 85% range of the rice-shrimp field, that is, Δt is between [110,150], and S1 mean When ≤ -23, it can be judged as a rice-shrimp field.

[0030] Embodiment 2: This embodiment is developed based on step 1 to step 4 of embodiment 1, and the same parts are not repeated here.

[0031] Step 4: Compare and judge based on key points and feature parameters, see Figure 8 , calculated from the mean of radar time series data and variance , through the normal rice-shrimp field , For comparison, when the mean and variance are within the 90% range of the rice-shrimp field, that is, At that time, it was identified as a rice-shrimp field.

[0032] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A rice-shrimp field identification method based on seasonal characteristics, characterized in that: The following steps are involved: Acquisition of time series data of paddy field radar; Radar time series data synthesis, missing value interpolation and filtering smoothing; Extract the key points of paddy field dynamics; Extraction of time series feature parameters; Compare and judge based on key points and feature parameters.

2. The rice-shrimp field identification method based on seasonal characteristics according to claim 1, characterized in that: The rice field radar time series feature extraction includes: the rice field state and the rice field radar time series feature value; The paddy field status includes: the paddy field status includes: the paddy field status S1 before rice transplanting, the rice growth status S2, and the paddy field status S3 after rice harvesting; The paddy field radar time series characteristic value is the average value of all radar backscatter coefficients in the paddy field at a certain moment.

3. The rice-shrimp field identification method based on seasonal characteristics according to claim 2, characterized in that: The paddy field states S1, S2, and S3 are represented by continuous integer digital codes, and their time series data are extracted according to the time periods of key change points.

4. The rice-shrimp field identification method based on seasonal characteristics according to claim 2, characterized in that: The radar time series data synthesis, interpolation and filtering smoothing include: synthesizing radar backscatter time series data at equal intervals to generate backscatter time series data with equal time intervals, then filling in missing values ​​through linear interpolation, and smoothing the radar backscatter signal time series data through the SG filtering algorithm.

5. The rice-shrimp field identification method based on seasonal characteristics according to claim 4, characterized in that: The radar backscatter time series data is obtained by performing boundary noise removal, speckle filtering, optimized terrain correction, and backscatter coefficient conversion on the VH polarization data of the Sentinel-1 GRD product.

6. The rice-shrimp field identification method based on seasonal characteristics according to claim 4, characterized in that: The slope of the paddy field radar backscatter coefficient curve is calculated using the difference method.

7. The rice-shrimp field identification method based on seasonal characteristics according to claim 1, characterized in that: The dynamic key point time of the paddy field includes: the time t1 of rice transplanting and the time t2 of rice harvesting; and is extracted according to the change in the slope of the rice growth state and the paddy field radar backscattering coefficient curve.

8. The rice-shrimp field identification method based on seasonal characteristics according to claim 1, characterized in that: The time series characteristic parameter extraction includes: extracting the mean value of the time series data of the paddy field radar backscatter coefficient and variance , the mean S1 of state S1 mean .

9. The rice-shrimp field identification method based on seasonal characteristics according to claim 7, characterized in that: The comparison and judgment based on key points and characteristic parameters include: comparison of key points and comparison of characteristic parameters, calculating the key point duration based on t1, and calculating the growth duration of the monitored object by comparing it with the growth duration of rice in the rice-shrimp field; calculating the mean, variance and mean statistical characteristic parameters of the radar time series data throughout the year, and judging whether it is a rice-shrimp field according to the growth duration and the range of the characteristic parameters by comparing with the key points and characteristic parameters of the rice-shrimp field.

10. The rice-shrimp field identification method based on seasonal characteristics according to claim 9, characterized in that: The key points and characteristic parameters are based on the duration of rice growth in the rice-shrimp field. When the growth time is within the 85% range of the rice-shrimp field, that is, Δt is between [110,150], and the characteristic parameters are used as reference. When the mean and variance are within the 90% range of the rice-shrimp field, that is, It was judged to be a rice-shrimp field.

Images