Shrimp-rice field information extraction method, electronic device and medium based on remote sensing satellite

Through Sentinel-2 images and land use data combined with the phenological characteristics of shrimp and rice fields, multi-time phase remote sensing image fusion and water index calculation are used to solve the problem of low accuracy of shrimp and rice fields in the existing technology, and high-precision and low-cost Extraction of information on shrimp and rice farming.

CN117576580BActive Publication Date: 2024-11-29GUANGDONG INST OF GEOLOGICAL SURVEYING & MAPPING CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202311541976.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-11-17
Publication Date
2024-11-29
Estimated Expiration
2043-11-17

AI Technical Summary

Technical Problem

The existing information extraction technology for growing shrimp rice fields based on Sentinel-2 remote sensing images has problems such as low spatial resolution, low calculation efficiency, complex calculation process, high calculation cost and low extraction accuracy, making it difficult to accurately obtain the spatio-temporal distribution information of shrimp rice fields.

Method used

Using Sentinel-2 images, land use data and phenological characteristics of shrimp and rice fields, the species of shrimp and rice fields are extracted by obtaining paddy field distribution range data, multi-time phase remote sensing image fusion, calculating water body index and segmenting images, and superimposing with paddy field distribution range to calculate the interception operation. Grow distribution information.

Benefits of technology

The accuracy and efficiency of information extraction of shrimp and rice fields are improved, and the problem of distinguishing shrimp and rice fields from other water bodies is solved, and high-precision and low-cost information extraction of shrimp and rice fields is achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117576580B_ABST
    Figure CN117576580B_ABST
Patent Text Reader

Abstract

A shrimp-rice field information extraction method, electronic device and medium based on remote sensing satellite, including: S1 obtains paddy field distribution range data of shrimp-rice field breeding information area to be extracted, and pre-processes the paddy field distribution range data to obtain the pre-processed paddy field distribution range; S2 obtains multi-temporal remote sensing images observed by Sentinel‑2 satellite corresponding to the shrimp-rice field breeding information area to be extracted; S3 fuses the multi-temporal remote sensing images using a median synthesis method to obtain a fused remote sensing image; S4 calculates a water body index based on the fused remote sensing image to obtain a result image, and then divides the result image into a water body area and a non-water body area; S5 superimposes the water body area obtained by S4 and the paddy field distribution range obtained by S1 to obtain the shrimp-rice field breeding distribution information. The present invention can effectively improve the accuracy of shrimp-rice field breeding information extraction by combining paddy field information in land use data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of satellite remote sensing, and in particular relates to a shrimp-rice field information extraction method based on a remote sensing satellite, electronic equipment and a medium. Background Art

[0002] The "shrimp-rice" farming model is an emerging rice planting model. It combines rice planting and crayfish (scientific name: Procambarus clarkii) farming in the same ecosystem. The "shrimp-rice" farming model can produce multiple types of food per unit area of ​​land, significantly improve the output efficiency of farmland, increase farmers' income, and is an effective way to achieve sustainable agricultural development and improve the ecological environment.

[0003] Accurately grasping the location, range, and area of ​​shrimp-rice fields used for "shrimp-rice farming" is of great significance for the promotion of the "shrimp-rice farming" model and the optimization of related industrial development. At present, the acquisition of information on shrimp-rice farming mainly relies on field surveys or statistical sampling methods to estimate its farming information. However, this method is difficult to accurately reflect the spatial distribution information of shrimp-rice fields in larger areas and finer spatial scales. How to accurately and timely obtain the spatiotemporal distribution information of shrimp-rice fields is a major technical challenge facing the "shrimp-rice farming" model in China's current promotion and industrial development decision-making.

[0004] Remote sensing technology has the ability to continuously monitor the surface in time and space, and can accurately reflect the spatial distribution of objects. It has become an important technical means to obtain surface cover information. Therefore, remote sensing images have been widely used in the extraction of objects such as farmland, cities and forests. The Sentinel-2 series of remote sensing satellites were first launched into space by the Copernicus program of the European Space Agency in 2015. The Sentinel-2 series of remote sensing satellites usually contain 12 bands, and compared with the Landsat series of satellites, the spatial resolution has been improved to about 10 meters, and the minimum period for repeated observation of the same area is only 5 days. And the remote sensing images obtained by this satellite can be obtained free of charge worldwide. The Sentinel-2 series of remote sensing satellites provide an important basic data source and guarantee for the refined extraction of shrimp and rice field breeding information. However, the existing technology for extracting shrimp and rice field breeding information based on Sentinel-2 remote sensing images still has problems such as low spatial resolution, low computational efficiency, complex calculation process, high computational cost, and low extraction accuracy. Summary of the invention

[0005] In view of the shortcomings of the above-mentioned prior art, the main purpose of the present invention is to utilize Sentinel-2 images, land use data and phenological characteristics of shrimp-rice fields to propose a refined extraction method for shrimp-rice field breeding information based on Sentinel-2 remote sensing images.

[0006] The technical solution of the present invention is as follows:

[0007] A method for extracting shrimp-rice field information based on remote sensing satellites comprises the following steps:

[0008] S1 obtains paddy field distribution range data of the shrimp-rice field breeding information area to be extracted, and preprocesses the paddy field distribution range data to obtain the preprocessed paddy field distribution range;

[0009] S2 obtains the multi-temporal remote sensing images observed by the Sentinel-2 satellite corresponding to the shrimp-rice field breeding information area to be extracted;

[0010] S3 uses the median synthesis method to fuse multi-temporal remote sensing images to obtain a fused remote sensing image;

[0011] S4 calculates the water body index based on the fused remote sensing image to obtain a result image, and then divides the result image into a water body area and a non-water body area;

[0012] S5 superimposes the water area obtained by S4 and the paddy field distribution range obtained by S1 and performs intersection operation to obtain the distribution information of shrimp and rice fields.

[0013] Furthermore, the preprocessing in S1 is: resampling and coordinate conversion of the paddy field distribution range data so that its spatial resolution and geographic coordinate system are consistent with the Sentinel-2 satellite image.

[0014] Furthermore, the multi-temporal remote sensing images in S2 are remote sensing images of all clouds with a cloud coverage of less than 5% observed by the Sentinel-2 satellite.

[0015] Furthermore, the median synthesis method formula is as follows:

[0016] ρ iM =median(ρ i1 , ρ i2 , ρ i3 …ρ it )

[0017] In the formula, ρ iM Represents the pixel value of the surface object in the area represented by a certain pixel after being synthesized on the i-th band of Sentinel-2; ρ i1 , ρ i2 , ρ i3 …ρ it They represent the pixel values ​​of the ground object in the i-th band of the remote sensing image at the time series nodes 1, 2, 3…n respectively; t represents the number of remote sensing images of qualified time series observations during the specified phenological period.

[0018] Furthermore, the resulting image is automatically segmented into water body areas and non-water body areas using an image segmentation algorithm.

[0019] Furthermore, based on the fused remote sensing image, the formula for calculating the water index is as follows:

[0020] MNDWI=(ρ Green -ρ SWIR1 ) / (ρ Green +ρ SWIR1 )

[0021] Where MNDWI represents the MNDWI water index calculated for the corresponding pixel of the ground object; ρ Green and ρ SWIR1 They represent the reflectance of the corresponding pixels of the ground objects in the Sentinel-2 image in the Green and SWIR1 bands respectively.

[0022] The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned shrimp-rice field information extraction method.

[0023] The present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, and when the computer program is executed by a processor, the above-mentioned shrimp-rice field information extraction method is implemented.

[0024] Technical effects of the present invention:

[0025] A shrimp-rice field breeding information extraction method was designed based on Sentinel-2 images. Compared with the commonly used Landsat images, Sentinel-2 satellite images have the advantages of high spatial resolution and short satellite repeated observation period, thus ensuring that the shrimp-rice field breeding information extraction has sufficient, sufficient, high-quality and high-precision basic data guarantee;

[0026] In the design of shrimp-rice field breeding information extraction technology, we fully utilized the unique phenological characteristics of the "shrimp-rice co-cultivation" breeding model, designed a system based on Sentinel-2 remote sensing data and combined it with paddy field information in land use data, which can effectively improve the accuracy of shrimp-rice field breeding information extraction. This solves the problem in the existing technology that it is difficult to effectively distinguish between shrimp-rice fields and lotus ponds and other land features by relying solely on the phenological and spectral characteristics of shrimp-rice fields in Sentinel-2 images.

[0027] Therefore, the technical solution proposed in the present invention has the advantages of low technical threshold for implementation, simple method principle, high calculation efficiency and easy promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings generally illustrate various embodiments by way of example and not limitation, and together with the description and claims, serve to illustrate the embodiments of the invention. Where appropriate, the same reference numerals are used throughout the drawings to refer to the same or similar parts. Such embodiments are illustrative and are not intended to be exhaustive or exclusive embodiments of the present apparatus or method.

[0029] Figure 1 A schematic diagram showing typical phenological characteristics of the "shrimp-rice co-cultivation" breeding model of the present invention is shown;

[0030] Figure 2 A schematic diagram of the method steps of the present invention is shown. DETAILED DESCRIPTION

[0031] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0032] like Figure 1-Figure 2 As shown, an embodiment of the present invention provides a method for extracting shrimp-rice field information based on Sentinel-2 remote sensing images, comprising the following steps:

[0033] (1) Acquisition of paddy field data: Obtain the paddy field distribution range data of the shrimp-rice field breeding area in the year to be extracted, and resample and transform the coordinates to make its spatial resolution and geographic coordinate system consistent with the Sentinel-2 satellite image. Shrimp-rice fields are essentially a type of paddy field with a relatively special utilization method. At present, high-precision paddy field data can be easily obtained globally and in China. In the specific implementation process, the distribution range of paddy fields can be obtained free of charge from the global land cover dataset, or from the land use change survey database of each local government each year. The resampling and coordinate transformation of paddy field distribution data can be completed through ArcGIS, QGIS and other software. For example, if you plan to extract the shrimp-rice field breeding distribution information of a certain area in 2020, you can obtain the paddy field distribution data that completely covers this area through the local government's annual land use change database.

[0034] (2) Remote sensing image selection: Through free distribution platforms for Sentinel-2 satellite images such as Google Earth Engine (GEE), obtain all remote sensing images with cloud coverage less than 5% observed by Sentinel-2 satellites during the crayfish farming period from December 1 to December 30 in the corresponding year for the shrimp-rice farming area to be extracted. For example, if you are going to extract the shrimp-rice farming distribution information of a certain area in 2020, you can use platforms such as GEE to obtain all images that completely cover this area and whose satellite observation time is between December 1 and December 30, 2020.

[0035] (3) Remote sensing image fusion: Using the median synthesis method, Band 3 (Green band, central wavelength 559.8nm), Band 4 (Red band, central wavelength 664.6nm), Band 8 (NIR band, central wavelength 832.8nm) and Band 11 (SWIR1 band, central wavelength 1613.7nm) of the multi-temporal remote sensing image are fused to obtain the fused remote sensing image, which is marked as Img1. Img1 contains a total of 4 bands, corresponding to the aforementioned Band 3, Band 4, Band 8 and Band 11. The formula of the median synthesis method is as follows:

[0036] ρ iM =median(ρ i1 , ρ i2 , ρ i3 …ρ it ) (1)

[0037] In the formula, ρ iM Represents the pixel value of the surface object in the area represented by a certain pixel after being synthesized on the i-th band of Sentinel-2; ρ i1 , ρ i2 , ρ i3 …ρ it They represent the pixel values ​​of the ground object in the i-th band of the remote sensing image at the time series nodes 1, 2, 3…n respectively; t represents the number of remote sensing images of qualified time series observations during the specified phenological period.

[0038] (4) Calculate the water index. Based on Img1, calculate the water index MNDWI (Modified Normalized Difference Water Index), and record the resulting image as MNDWI. The calculation formula of MNDWI is as follows:

[0039] MNDWI=(ρ Green -ρ SWIR1 ) / (ρ Green +ρSWIR1 ) (2)

[0040] Where MNDWI represents the MNDWI water index calculated for the corresponding pixel of the ground object; ρ Green and ρ SWIR1 They represent the reflectance (pixel value) of the corresponding pixels of the ground objects in the Sentinel-2 image in the Green (Band 3) and SWIR1 (Band 11) bands respectively.

[0041] (5) Extracting water bodies. Based on the MNDWI image calculated in step (4), the Otsu method is used to automatically segment the MNDWI image into water and non-water areas. The areas identified as water bodies are marked as WATER, and other non-water areas are directly removed.

[0042] (6) Extracting shrimp-rice field cultivation information. The WATER image obtained in step (5) and the paddy field range image in step (1) are superimposed and intersected to obtain the shrimp-rice field cultivation distribution information in a certain year.

[0043] The quality information such as the cloud coverage area ratio of the remote sensing image described in the step (2) of selecting the remote sensing image can be obtained by querying the metadata of the remote sensing image.

[0044] The step (3) remote sensing image fusion: the remote sensing image is stored in the computer in the form of a multi-band two-dimensional matrix. One band corresponds to a two-dimensional matrix, and one element of the two-dimensional matrix corresponds to a pixel on the remote sensing image, representing a rectangular area in the real world. For example, the value of an element in the two-dimensional matrix used to store a band (such as the Red band) on the Sentinel-2 image represents the reflectivity of a 10m×10m area in the Red band. Therefore, the fusion calculation of images and bands can be implemented with the help of existing mature software platforms or technologies such as GEE, ArcGIS, Python programming technology, etc., based on the formula provided by this technical solution.

[0045] The median synthesis function in the remote sensing image fusion in step (3) is based on ρ i1 , ρ i2 , ρ i3 …ρ it In the specific implementation, firstly, ρ i1 , ρ i2 , ρ i3 …ρ itThe values ​​are arranged in descending or ascending order, and then the number in the middle is taken as the median. In specific implementation, the calculation of the median can be completed by directly calling the functions provided by existing mature software platforms or technologies such as GEE, ArcGIS, and Python programming technology.

[0046] The median synthesis in the remote sensing image fusion of step (3) and obtaining the synthetic image Img1 is to eliminate the influence of noise such as clouds and fog on the quality of remote sensing images and improve the accuracy of shrimp and rice field breeding information extraction.

[0047] The Otsu method described in step (5) is a method for binarizing an image, and its calculation ideas include: ① Calculate the histogram of the NDVI and MNDWI images to obtain the number of pixels at each gray level; ② For each possible threshold, divide the image into two parts according to it; ③ Calculate the number of pixels and the average gray value of each part after segmentation; ④ Calculate the inter-class variance. ⑤ Select the gray value that maximizes the inter-class variance as the threshold. The Otsu method is not an original method of the present invention, and this scheme does not elaborate on the Otsu algorithm in detail. In the specific implementation process, the Otsu algorithm can be completed with the help of functions provided by existing mature software platforms or technologies such as GEE and Python programming technology.

[0048] The basic principle behind the information technology for extracting the distribution of shrimp and rice fields in step (6) is determined by the unique phenological characteristics of the "shrimp and rice co-cultivation" farming model. The specific principles include: ① Between October and May each year, the shrimp and rice fields are in the state of lobster farming, and the rice fields are filled with water, and the water depth is usually higher than 50 cm. ② During this phenological period, the shrimp and rice fields are in the state of water bodies, and will show similar spectral reflectance characteristics in the Sentinel-2 image, just like water bodies such as ponds, lakes, reservoirs and rivers. ③ For other ordinary paddy fields, they are generally in a dry farming state during the dry season in winter, showing completely different spectral characteristics from shrimp and rice fields. Therefore, through the Sentinel-2 synthetic image of the lobster farming period in December, we can extract the shrimp and rice fields, ponds, lakes, reservoirs and rivers and other water bodies in the region from the remote sensing image. The result data extracted by step (5) includes various water bodies such as shrimp and rice fields, ponds, reservoirs, lakes and rivers. ④ Only those rice fields that still show water body characteristics in the winter dry season are shrimp-rice fields. Therefore, by intersecting the two sets of water body and paddy field range obtained in step (5), we can obtain high-precision shrimp-rice field breeding information. In the specific implementation process, the functions or functions provided by existing mature software platforms or technologies such as GEE, ArcGIS, Python programming technology, etc. can be used to intersect the results of the above two steps to obtain refined shrimp-rice field breeding information.

[0049] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for extracting shrimp-rice field information based on remote sensing satellites, characterized in that: The following steps are involved: S1 obtains paddy field distribution range data of the shrimp-rice field breeding information area to be extracted, and preprocesses the paddy field distribution range data to obtain the preprocessed paddy field distribution range; S2 obtains the multi-temporal remote sensing images observed by the Sentinel-2 satellite corresponding to the shrimp-rice field breeding information area to be extracted; S3 uses the median synthesis method to fuse multi-temporal remote sensing images to obtain a fused remote sensing image; S4 calculates the water index based on the fused remote sensing image to obtain a result image, and then divides the result image into a water area and a non-water area; specifically: using an image segmentation algorithm to automatically divide the result image into a water area and a non-water area; S5 performs superposition and intersection operations on the water area obtained by S4 and the paddy field distribution range obtained by S1 to obtain the distribution information of shrimp and rice fields; The median synthesis method formula is as follows: r iM =median(ρ i1 ,r i2 ,r i3 …r it ) In the formula, ρ iM It represents the pixel value of the surface object in the area represented by a certain pixel after being synthesized on the i-th band of Sentinel-2; ρ i1 , ρ i2 , ρ i3 …ρ it Respectively represent the pixel value of the surface object on the i-th band of the remote sensing image at the time series 1, 2, 3..n...0 point; t represents the number of remote sensing images of qualified time series observations during the specified phenological period; The preprocessing in S1 is: resampling and coordinate conversion of the paddy field distribution range data to make the spatial resolution and geographic coordinate system of the paddy field distribution range data consistent with the Sentine1-2 satellite image; The multi-temporal remote sensing images in S2 are remote sensing images with cloud coverage less than 5% observed by the Sentinel-2 satellite; Based on the fused remote sensing image, the formula for calculating the water body index is as follows: MNDWI=(ρ Green -r SWIR1 ) / (ρ Green +r SWIR1 ) Where MNDWI represents the MNDWI water index calculated for the corresponding pixel of the ground object; ρ Green and ρ SWIR1 They represent the reflectance of the corresponding pixels of the ground objects in the Sentinel-2 image in the Green and SWIR1 bands respectively.

2. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the shrimp-rice field information extraction method according to claim 1.

3. A computer-readable storage medium, characterized in that: It stores a computer program, and when the computer program is executed by a processor, the shrimp-rice field information extraction method as claimed in claim 1 is implemented.

Citation Information

Patent Citations

  • Rice planting area extraction method based on multi-source data

    CN114821362A