Multi-source remote sensing image data fusion processing method

Abstract

The invention discloses a multi-source remote sensing image data fusion processing method, and the method comprises the following steps: carrying out the preprocessing of Sentinel-1 data and Sentinel-2 data, and guaranteeing that the Sentinel-2 data and the Sentinel-1 data are in the same projection coordinate system; extracting wave band data in the preprocessed Sentinel-1 data and the preprocessed Sentinel-2 data, and synthesizing the extracted wave band data; generating a new data set containing the fused Sentinel-1 data and the Sentinel-2 data, and generating a new data set containing the fused Sentinel-1 data and Sentinel-2 data; according to the method, the SAR data (Sentinel-1 data) and the optical data (Sentinel-2 data) are fused, the complementary advantages of the SAR data and the optical data are fully utilized, the precision and the effect of image analysis are improved, a new data set containing the merged and fused wavebands is finally generated, and richer data information is provided for subsequent remote sensing image analysis and application.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing data processing, and in particular to a multi-source remote sensing image data fusion processing method. Background Art

[0002] Multi-source remote sensing image data fusion refers to the integration of data from different sensors (such as optical, radar and other types of remote sensing data) to obtain more comprehensive and accurate information. This technology has a wide range of applications in the fields of geographic information systems (GIS), environmental monitoring, agriculture, urban planning and disaster management. Among them, the most commonly used remote sensing impact data are Sentinel-1 data and Sentinel-2 data.

[0003] Among them, Sentinel-1 is a C-band synthetic aperture radar (SAR) satellite of the European Space Agency (ESA), which can obtain surface information under any weather conditions. Sentinel-1 data mainly includes VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model) data, and Sentinel-2 is an optical imaging satellite of the European Space Agency, providing high-resolution multispectral images. Sentinel-2 data mainly includes visible light and near-infrared bands (bands 2, 3, 4, 8, namely R, G, B and NIR).

[0004] In the preprocessing steps of multi-source remote sensing image data, most of them are limited to the processing and application of single sensor data, specifically:

[0005] Many studies focus only on the preprocessing and application of Sentinel-1 data, such as surface change detection and disaster monitoring through SAR data. These studies usually involve processing steps such as clipping, thermal noise removal, orbit correction, radiometric calibration, speckle filtering, terrain correction and decibelization, but ultimately only the processed SAR data are used for analysis and application.

[0006] Similarly, many studies focus only on the preprocessing and application of Sentinel-2 data, such as vegetation index calculation and land cover classification through multispectral data. These studies usually involve processing steps such as clipping and band synthesis, but ultimately only use the processed optical data for analysis and application.

[0007] Therefore, the above-mentioned processing and application is limited to single sensor data, resulting in the inability to fully utilize the complementary advantages of multi-source data, limiting the breadth and accuracy of applications. Summary of the invention

[0008] The purpose of the present invention is to provide a multi-source remote sensing image data fusion processing method to solve the technical problem that the prior art is limited to the processing and application of single sensor data, resulting in the inability to fully utilize the complementary advantages of multi-source data, limiting the breadth and accuracy of application.

[0009] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0010] A multi-source remote sensing image data fusion processing method comprises the following steps:

[0011] Step 100, pre-process the Sentinel-1 data and the Sentinel-2 data respectively, and ensure that the Sentinel-2 data is in the same projection coordinate system as the Sentinel-1 data;

[0012] Step 200: extracting band data from the preprocessed Sentinel-1 data and Sentinel-2 data, and synthesizing the extracted band data;

[0013] Step 300: Generate a new data set including the fusion of Sentinel-1 data and Sentinel-2 data.

[0014] As a preferred solution of the present invention, in step 100, the steps of preprocessing the Sentinel-1 data are as follows:

[0015] Modeling the background thermal noise of the Sentinel-1 satellite sensor and removing it from the raw image data;

[0016] Apply the radiation calibration coefficients corresponding to the Sentinel-1 satellite sensor to convert the DN values ​​in the Sentinel-1 data into radiation quantities;

[0017] Apply filters to filter out speckle while preserving the edges of Sentinel-1 data;

[0018] Use digital elevation model data to correct terrain effects in Sentinel-1 data;

[0019] Convert the amplitude data of Sentinel-1 data to a logarithmic scale to obtain the decibel scale of Sentinel-1 data;

[0020] The Sentinel-1 data is cropped so that the Sentinel-1 data is confined to a specific area.

[0021] As a preferred solution of the present invention, in step 100, the implementation method of preprocessing the Sentinel-2 data is:

[0022] Using a multi-scale mosaic method, the projection of the Sentinel-2 dataset is unified into the WGS84 coordinate system, so that the Sentinel-2 data and the Sentinel-1 data are in the same projection coordinate system;

[0023] The band extraction method is used to extract four bands, namely red (R), green (G), blue (B) and near infrared (NIR), from Sentinel-2 data;

[0024] The Sentinel-2 data is cropped to extract data for specific geographic areas.

[0025] As a preferred solution of the present invention, the Sentinel-2 data after clipping is data that only retains the positioning station data as the center point;

[0026] The Sentinel-1 data after clipping is data that only retains the positioning station data as the center point;

[0027] The cropped Sentinel-2 data and the cropped Sentinel-1 data have the same size.

[0028] As a preferred solution of the present invention, the Sentinel-1 data includes VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model).

[0029] As a preferred solution of the present invention, the waveform data extracted from the Sentinel-1 data includes VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model);

[0030] The band data extracted from the Sentinel-1 data and the waveform data extracted from the Sentinel-2 data are merged into a 7-band labeled image file format dataset in the order of VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model), R (red), G (green), B (blue) and NIR (infrared).

[0031] As a preferred solution of the present invention, in step 300, the Sentinel-1 data and the Sentinel-2 data are used to create multiple new data sets including band merging and fusion.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention provides a method for fusing Sentinel-1 and Sentinel-2 data. By fusing SAR data (Sentinel-1 data) and optical data (Sentinel-2 data), the complementary advantages of the two are fully utilized to improve the accuracy and effect of image analysis, and finally a new data set containing the merged and fused bands is generated, thereby providing richer data information for subsequent remote sensing image analysis and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0035] Figure 1 A schematic diagram of a data fusion processing method according to an embodiment of the present invention; DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] like Figure 1 As shown, the present invention provides a multi-source remote sensing image data fusion processing method, including geometric correction, coordinate conversion, data alignment and registration, and the implementation of a fusion algorithm, specifically comprising the following steps:

[0038] Step 100: Preprocess the Sentinel-1 data and the Sentinel-2 data respectively, and ensure that the Sentinel-2 data is in the same projection coordinate system as the Sentinel-1 data.

[0039] The preprocessing steps for Sentinel-1 and Sentinel-2 data include thermal noise removal, orbit correction, radiation calibration, speckle filtering, terrain correction, decibelization, band synthesis, etc. Integrate the automated process of multi-source data processing and fusion to improve processing efficiency

[0040] In step 100, the steps for preprocessing the Sentinel-1 data are:

[0041] (1) Model the background thermal noise of the Sentinel-1 satellite sensor and remove the thermal noise from the raw image data. In this step, the image quality is improved by removing the thermal noise generated by the sensor itself (such as the signal caused by the temperature change inside the sensor), which improves the accuracy and reliability of the data, especially when conducting remote sensing observations under low temperature conditions.

[0042] (2) Apply the radiation calibration coefficients corresponding to the Sentinel-1 satellite sensor to convert the DN values ​​in the Sentinel-1 data into radiation quantities, which usually includes converting the DN values ​​into radiation brightness or surface reflectivity, that is, converting the original digital values ​​of the image into actual physical quantities (such as radiation intensity or reflectivity) to ensure that the data are comparable between different times and sensors for accurate analysis and comparison.

[0043] (3) Apply filters for speckle filtering to remove speckle noise from Synthetic Aperture Radar (SAR) images, especially noise patterns commonly found in SAR images, while preserving Sentinel-1 data edges and details.

[0044] The filter includes a Wiener filter, a Lee filter or other statistical filters, which reduces the influence of noise by analyzing the local statistical characteristics in the image, thereby improving the image quality and the accuracy of target detection.

[0045] (4) Use digital elevation model data to correct the terrain effect in Sentinel-1 data. This step takes into account the impact of terrain undulations on image geometry to correct remote sensing images. The correction includes adjusting the image geometry to make it consistent with the terrain information and improving the geographic accuracy of the image. The DEM data is retained as one of the bands in the corrected Sentinel-1 data.

[0046] (5) The amplitude data of Sentinel-1 data is converted into a logarithmic scale to obtain the decibel scale of Sentinel-1 data, that is, the intensity values ​​in the Sentinel-1 data are converted into a decibel (dB) scale for quantification and analysis, so that the intensity values ​​in the image are more consistent with actual perception and easier to analyze and interpret.

[0047] (6) Crop the Sentinel-1 data so that they are confined to a specific area. Crop the image according to the defined geographic range or coordinate system, remove the unnecessary parts, and only keep the data of the positioning station with the center point of 1000*1000 and 320*320, so as to improve the efficiency and accuracy of subsequent analysis.

[0048] In step 100, the implementation method of preprocessing the Sentinel-2 data is as follows:

[0049] (1) Using the multi-scale mosaic method, the projection of the Sentinel-2 dataset is unified into the WGS84 coordinate system, so that the Sentinel-2 data and the Sentinel-1 data are in the same projection coordinate system.

[0050] The Multi-Size Mosaic method in the SNAP software was used to unify the projection of the Sentinel-2 dataset into the WGS84 coordinate system, which was used to synthesize images or datasets of different resolutions, ensuring that the Sentinel-2 data was in the same projection coordinate system as the Sentinel-1 data, in order to facilitate subsequent processing and analysis and maintain data consistency and accuracy.

[0051] (2) Use the band extraction method to extract four bands: red (R), green (G), blue (B) and near-infrared (NIR) from the Sentinel-2 data.

[0052] The band extraction (Subset) method in the SNAP software is used to extract specific areas or bands from the Sentinel-2 data for further analysis and processing, reducing the amount of data processing and improving processing efficiency.

[0053] (3) Cropping of Sentinel-2 imagery to extract data for specific geographic areas.

[0054] The Clip method in the SNAP software is used to clip the Sentinel-2 data to extract the process of a specific geographical area, retaining only the data of the positioning station with the center point of 1000*1000 and 320*320, removing irrelevant areas to ensure the accuracy and relevance of the data.

[0055] Therefore, the cropped Sentinel-2 data only retains the positioning station data as the center point; the cropped Sentinel-1 data only retains the positioning station data as the center point; the cropped Sentinel-2 data and the cropped Sentinel-1 data have the same size.

[0056] Step 200: extract band data from the preprocessed Sentinel-1 data and Sentinel-2 data, and synthesize the extracted band data.

[0057] Sentinel-1 data contains VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model). Waveform data extracted from Sentinel-1 data includes VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model).

[0058] The band data extracted from the Sentinel-1 data and the waveform data extracted from the Sentinel-2 data are merged into a 7-band labeled image file format dataset in the order of VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model), R (red), G (green), B (blue) and NIR (infrared).

[0059] Step 300: Generate a new data set containing the fused Sentinel-1 data and Sentinel-2 data, wherein the created new data set containing the fused data provides richer data information for subsequent remote sensing image analysis and application. This data set has higher accuracy and information content, and is suitable for a variety of remote sensing application fields, such as surface change detection, environmental monitoring, agricultural management, urban planning, etc.

[0060] In step 300, multiple new datasets containing merged bands are created by combining Sentinel-1 data and Sentinel-2 data.

[0061] The new data sets finally obtained are 1000*1000 and contain the new data sets after the bands are merged and fused, and 320*320 and contain the new data sets after the bands are merged and fused.

[0062] The fusion process of the above-mentioned Sentinel-1 data and Sentinel-2 data can be implemented through Python code, and GPT is called through the command line to run the SNAP flowchart, thus completing the automation of the entire process.

[0063] Therefore, this embodiment proposes a method for fusing Sentinel-1 and Sentinel-2 data. By fusing SAR data (Sentinel-1 data) and optical data (Sentinel-2 data), the complementary advantages of the two are fully utilized to improve the accuracy and effect of image analysis. Finally, a new data set containing the merged and fused bands is generated to provide richer data information for subsequent remote sensing image analysis and application. The technical implementation and advantages of the fused data set in specific applications, including surface change detection, environmental monitoring, agricultural management, urban planning, etc., and technical solutions to improve data processing and fusion efficiency, including the application of high-performance computing technology and algorithm optimization, and the design and implementation of an automated processing framework.

[0064] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A multi-source remote sensing image data fusion processing method, characterized in that: The following steps are involved: Step 100, pre-process the Sentinel-1 data and the Sentinel-2 data respectively, and ensure that the Sentinel-2 data is in the same projection coordinate system as the Sentinel-1 data; Step 200: extracting band data from the preprocessed Sentinel-1 data and Sentinel-2 data, and synthesizing the extracted band data; Step 300: Generate a new data set including the fusion of Sentinel-1 data and Sentinel-2 data.

2. The multi-source remote sensing image data fusion processing method according to claim 1, characterized in that: In step 100, the steps of preprocessing the Sentinel-1 data are as follows: Modeling the background thermal noise of the Sentinel-1 satellite sensor and removing it from the raw image data; Apply the radiation calibration coefficients corresponding to the Sentinel-1 satellite sensor to convert the DN values ​​in the Sentinel-1 data into radiation quantities; Apply filters to filter out speckle while preserving the edges of Sentinel-1 data; Use digital elevation model data to correct terrain effects in Sentinel-1 data; Convert the amplitude data of Sentinel-1 data to a logarithmic scale to obtain the decibel scale of Sentinel-1 data; The Sentinel-1 data is cropped so that the Sentinel-1 data is confined to a specific area.

3. The multi-source remote sensing image data fusion processing method according to claim 2 is characterized in that: In step 100, the implementation method of preprocessing the Sentinel-2 data is: Using a multi-scale mosaic method, the projection of the Sentinel-2 dataset is unified into the WGS84 coordinate system, so that the Sentinel-2 data and the Sentinel-1 data are in the same projection coordinate system; The band extraction method is used to extract four bands, namely red (R), green (G), blue (B) and near infrared (NIR), from Sentinel-2 data; The Sentinel-2 data is cropped to extract data for specific geographic areas.

4. The multi-source remote sensing image data fusion processing method according to claim 3 is characterized in that: The Sentinel-2 data after clipping is data that only retains the positioning station data as the center point; The Sentinel-1 data after clipping is data that only retains the positioning station data as the center point; The cropped Sentinel-2 data and the cropped Sentinel-1 data have the same size.

5. A multi-source remote sensing image data fusion processing method according to claim 3 or 4, characterized in that: The Sentinel-1 data includes VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model).

6. The multi-source remote sensing image data fusion processing method according to claim 5, characterized in that: Extracting waveform data including VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model) from the Sentinel-1 data; The band data extracted from the Sentinel-1 data and the waveform data extracted from the Sentinel-2 data are merged into a 7-band labeled image file format dataset in the order of VH (vertical-horizontal polarization), VV (vertical-vertical polarization) and DEM (digital elevation model), R (red), G (green), B (blue) and NIR (infrared).

7. The multi-source remote sensing image data fusion processing method according to claim 6, characterized in that: In step 300, the Sentinel-1 data and the Sentinel-2 data are used to create multiple new data sets including band merging and fusion.

Images