Multi-source remote sensing optical data intelligent access and processing method, electronic equipment and storage medium

Abstract

The invention discloses a multi-source remote sensing optical data intelligent access and processing method, electronic equipment and a storage medium, relates to the technical field of remote sensing data processing, and solves the technical problems that multi-source remote sensing optical data access automation is insufficient, quality inspection and storage management are disordered, processing triggering lacks accurate judgment logic, and there is no scientific optimization mechanism for multiple qualified images. Multi-source public remote sensing optical original data are automatically grabbed, after qualified data are screened through quality inspection, a data and coordinate separation strategy is adopted for storage and tagging, an index is established, and then the full coverage condition of the data on a target area is judged through spatial intersection operation so as to trigger a processing flow. And finally calling qualified data to execute processing and outputting a remote sensing product, thereby realizing full-process automatic efficient processing. According to the invention, automation and intelligentization of the whole process from data acquisition to final product generation are realized, and the efficiency and accuracy of remote sensing data processing are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of remote sensing data processing technology, specifically to a method for intelligent access and processing of multi-source remote sensing optical data, an electronic device, and a storage medium. Background Technology

[0002] With the rapid development of remote sensing technology, the availability of multi-source public remote sensing data sources such as the Landsat series, Sentinel series, MODIS, and Gaofen series is becoming increasingly abundant. Remote sensing optical data is being used more and more widely in fields such as geographic information acquisition, environmental monitoring, disaster early warning, agricultural monitoring, and urban planning, and the scale of data is also showing explosive growth.

[0003] Currently, while existing technologies possess basic capabilities for remote sensing data acquisition and processing—for example, some data sources support manual downloading or simple batch capture—and some scattered quality inspection tools and data processing algorithms exist, several significant shortcomings remain: First, the automation level of data access is low, relying heavily on manual parameter configuration and data filtering, and the interface compatibility between different data sources is poor, making it difficult to achieve unified and efficient capture of multi-source data; second, the quality inspection process is incomplete, focusing only on basic indicators such as file size and format, lacking in-depth inspection dimensions such as spectral consistency and data integrity, leading to unqualified data flowing into subsequent stages and affecting product quality; third, data storage management is chaotic. The existing remote sensing data processing workflow suffers from several problems. First, it often uses a mixed storage method of data files, metadata, and coordinate information, lacking a standardized labeling system and efficient indexing mechanism. This results in time-consuming and low-utilization retrieval of massive amounts of data. Second, the processing triggering mechanism is not intelligent, failing to automatically trigger processing flows based on accurate spatial coverage. This leads to either blindly initiating processing, wasting resources, or missing effective processing opportunities due to insufficient coverage. Furthermore, when multiple qualified images exist for the same area, there is a lack of scientific optimization models, relying heavily on manual experience for selection, resulting in low processing efficiency and unstable product quality. Third, the multi-source data collaboration capability is weak, making it difficult to achieve intelligent stitching and fusion of incompletely covered data, and failing to fully leverage the advantages of different data sources. These issues result in long processing cycles (often requiring several days), excessive manual intervention, low efficiency, and difficulty in guaranteeing the final product quality, failing to meet the urgent needs of various fields for rapid, accurate, and high-quality remote sensing data processing. Summary of the Invention

[0004] To address the technical problems of insufficient automation in multi-source remote sensing optical data access, chaotic quality inspection and storage management, lack of precise judgment logic for processing triggers, and absence of a scientific optimization mechanism for multiple qualified images, this invention provides an intelligent access and processing method for multi-source remote sensing optical data, including:

[0005] S1. Automatic Acquisition: Based on the preset task list and scheduling strategy, automatically acquire raw remote sensing optical data from multiple public remote sensing data sources;

[0006] S2. Data Access and Quality Inspection: Receive the raw remote sensing optical data, perform quality inspection on it, filter out the data that passes the quality inspection and output it.

[0007] S3. Data storage: Receive the qualified data, store it using a storage strategy that separates data files and coordinate information, and label each data unit with one or more tags including: data source, collection time and quality level. Establish a data index with "administrative division-time" as the basic dimension, sort the data quality, and output the stored structured data with tags and indexes.

[0008] S4. Processing trigger judgment: Receive the structured data, parse the metadata in it to obtain the spatial coverage of the data, perform spatial intersection calculation with the preset target area administrative division, if the calculation result shows that the access data corresponding to the structured data has achieved full coverage of the target area, then automatically trigger the subsequent data processing process and output the processing task instruction.

[0009] The spatial intersection operation uses a spatial analysis method based on vector polygons or raster data to calculate the overlap between the data coverage area and the target area.

[0010] S5. Data Processing: Receive the processing task instruction, call the corresponding qualified data in the structured data in S3, execute the data processing flow, and output the remote sensing processing product.

[0011] Furthermore, in S1, the plurality of publicly disclosed remote sensing data sources include, but are not limited to, at least two of the following: Landsat series, Sentinel series, MODIS, Gaofen series, and Planet data sources;

[0012] The automatic capture includes at least one of the following: capture period, time range, spatial range, cloud cover threshold, and data resolution.

[0013] Furthermore, S1 specifically includes:

[0014] S1.1 Automatic data capture initialization, preset capture period, data source list, time range, spatial range, cloud cover threshold parameters, and output preset parameter set;

[0015] S1.2 Construct and send an API query request, dynamically constructing a query request that conforms to the API interface specification of the target data source based on the preset parameter set;

[0016] S1.3. Based on the query request, parse the API response and filter the results to form a list of original data to be downloaded;

[0017] S1.4 Raw data download: Traverse the list of raw data to be downloaded and initiate a download request to obtain the remote sensing optical raw data;

[0018] S1.5 Metadata Extraction and Task Log Recording: Extract key information from the downloaded remote sensing optical raw data and generate an execution log.

[0019] Furthermore, S2 specifically includes:

[0020] S2.1. Check the integrity of the compressed file to verify whether the decompression process is normal.

[0021] S2.2 Required file completeness check: Check whether the unzipped folder contains all required files;

[0022] S2.3 Data size check: Verify that the data file size is within a reasonable range;

[0023] S2.4 File naming correctness check: Verify whether the file naming conforms to the preset naming convention.

[0024] Furthermore, in S3, the storage strategy of separating the data file and coordinate information is specifically as follows: the original image data is stored in a distributed file system or object storage, and the coordinate information, metadata and quality inspection results are stored in a spatial database;

[0025] The quality levels are based on administrative divisions and time constraints, ranking data quality with cloud coverage <5% as excellent, <10% as good, and <20% as average.

[0026] Furthermore, S4 specifically includes:

[0027] S4.1: Parse the spatial coverage of the currently accessed data from the stored metadata;

[0028] S4.2: Perform a spatial intersection operation between the spatial coverage area and the preset target area to determine the administrative region to which it belongs; the formula for the spatial intersection operation is:

[0029] Containment relationship: Area(I ∩ C) = Area(I)

[0030] Comparison:

[0031] Where I represents access data, C represents administrative division data, and Area() represents the area calculation function;

[0032] S4.3: Determine whether the data coverage area completely includes or equals the target area, using the following formula:

[0033]

[0034] in, represent , represents the effective area of ​​the nth image, that is, the access data without cloud coverage, and ε represents the allowable uncovered area. For example, if the allowable uncovered area is 1%, then ε is 0.01.

[0035] S4.4: If the data is determined to be fully covered, a processing task will be automatically generated and subsequent data processing will be triggered. If the data is determined to be partially covered, it will be marked as data to be stitched together and will wait to be combined with other data to form full coverage before the processing process is triggered. For partially covered data, the system will automatically search for stored data from adjacent time phases or adjacent orbits for stitching. When the stitched data meets the full coverage requirements, the processing process will be automatically triggered.

[0036] S4.5: Construct an evaluation index system, including six major indicators: cloud coverage ratio, spatial resolution, signal-to-noise ratio, temporal suitability, data source priority, and radiation consistency.

[0037] S4.6: Each image is scored using a comprehensive scoring model.

[0038] Wherein, Score is the overall score, c is the cloud coverage ratio, R is the spatial resolution, SNR is the signal-to-noise ratio, T is the temporal suitability, S is the data source priority, and Rad is the radiation consistency. Assigning weights to each indicator. This is the normalization function for each indicator;

[0039] S4.7: Select the image with the highest overall score as the priority for processing.

[0040] Furthermore, in S5, the data processing flow specifically includes:

[0041] S5.1: Radiometric calibration, which converts the raw digital quantization values ​​of an image into physically meaningful surface reflectance or radiance values;

[0042] S5.2: Atmospheric correction, eliminating the effects of atmospheric scattering and absorption on remote sensing images;

[0043] S5.3: Geometric correction, eliminating image geometric distortion caused by sensor orientation and terrain undulation;

[0044] S5.4: Image enhancement, which improves image quality through histogram equalization and filtering methods;

[0045] S5.5: Data fusion, which involves fusing multi-source, multi-resolution remote sensing data;

[0046] S5.6: Thematic extraction, using machine learning or deep learning algorithms for land cover classification, change detection, and target recognition.

[0047] An electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the access and processing method as described in any of the preceding claims.

[0048] A computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the access and processing method as described in any of the preceding claims.

[0049] The beneficial effects of this invention are as follows: This invention automates and intelligently processes multi-source remote sensing optical data, from automatic acquisition, multi-dimensional quality inspection, separate storage and structured label management, to precise processing triggering based on spatial computation, intelligent image selection based on multi-index comprehensive evaluation, and standardized data processing. This not only significantly reduces manual intervention and shortens the data processing cycle from several days to several hours, but also improves data retrieval efficiency through separate storage and multi-dimensional labeling, avoids wasting resources through spatial coverage judgment, ensures input data quality based on multi-index selection models, and continuously improves system performance through adaptive weight optimization and multi-source data collaboration mechanisms. At the same time, a comprehensive quality inspection system and anomaly handling mechanism ensure the reliability and accuracy of the final product. It also has high scalability and flexible adaptability, effectively meeting the needs of multiple fields such as geographic information acquisition and environmental monitoring for efficient and high-quality remote sensing data processing. Attached Figure Description

[0050] Figure 1 This is a flowchart of the present invention;

[0051] Figure 2 This is a logic diagram of S4 of the present invention. Detailed Implementation

[0052] The technical solution of the present invention will be further described below with reference to embodiments, but it is not limited thereto. Any modifications or equivalent substitutions to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be covered within the protection scope of the present invention. In the following embodiments, the process equipment or apparatus not specifically specified are all conventional equipment or apparatus in the art. Unless otherwise specified, the raw materials used in the embodiments of the present invention are all commercially available; unless otherwise specified, the technical means used in the embodiments of the present invention are all conventional means well known to those skilled in the art.

[0053] Example 1, combined with Figure 1 and Figure 2This embodiment describes a method for intelligent access and processing of multi-source remote sensing optical data, including:

[0054] S1. Automatic Acquisition: Based on the preset task list and scheduling strategy, automatically acquire raw remote sensing optical data from multiple public remote sensing data sources;

[0055] S2. Data Access and Quality Inspection: Receive the raw remote sensing optical data, perform quality inspection on it, filter out the data that passes the quality inspection and output it.

[0056] S3. Data storage: Receive the qualified data, store it using a storage strategy that separates data files and coordinate information, and label each data unit with one or more tags including: data source, collection time and quality level. Establish a data index with "administrative division-time" as the basic dimension, sort the data quality, and output the stored structured data with tags and indexes.

[0057] S4. Processing trigger judgment: Receive the structured data, parse the metadata in it to obtain the spatial coverage of the data, perform spatial intersection calculation with the preset target area administrative division, if the calculation result shows that the access data corresponding to the structured data has achieved full coverage of the target area, then automatically trigger the subsequent data processing process and output the processing task instruction.

[0058] The spatial intersection operation uses a spatial analysis method based on vector polygons or raster data to calculate the overlap between the data coverage area and the target area.

[0059] S5. Data Processing: Receive the processing task instruction, call the corresponding qualified data in the structured data in S3, execute the data processing flow, and output the remote sensing processing product.

[0060] Specifically, this invention achieves fully automated processing with minimal human intervention by automatically capturing, screening, separating, storing, and indexing multi-source remote sensing optical data, and by using spatial intersection operations to determine the coverage of the target area. Full coverage triggers the processing flow. This improves the efficiency of data management and processing, and ensures the quality and accuracy of the final remote sensing processing product.

[0061] pass Figure 2 The process of S4 is as follows: First, the metadata is parsed to obtain the spatial coverage. After spatial intersection operation and image quality sorting, it is determined whether the data completely covers the target area. If it does not cover, it is marked as data to be stitched and adjacent data is searched for and stitched. After stitching, the coverage is verified again. Once the coverage is confirmed, the subsequent processing process is triggered.

[0062] In S1, the plurality of publicly disclosed remote sensing data sources include, but are not limited to, at least two of the following: Landsat series, Sentinel series, MODIS, Gaofen series, and Planet data sources;

[0063] The automatic capture includes at least one of the following: capture period, time range, spatial range, cloud cover threshold, and data resolution.

[0064] S1 specifically includes:

[0065] S1.1 Automatic data capture initialization, preset capture period, data source list, time range, spatial range, cloud cover threshold parameters, and output preset parameter set;

[0066] S1.2 Construct and send an API query request, dynamically constructing a query request that conforms to the API interface specification of the target data source based on the preset parameter set;

[0067] S1.3. Based on the query request, parse the API response and filter the results to form a list of original data to be downloaded;

[0068] S1.4 Raw data download: Traverse the list of raw data to be downloaded and initiate a download request to obtain the remote sensing optical raw data;

[0069] S1.5 Metadata Extraction and Task Log Recording: Extract key information from the downloaded remote sensing optical raw data and generate an execution log.

[0070] S2 specifically includes:

[0071] S2.1. Check the integrity of the compressed file to verify whether the decompression process is normal.

[0072] S2.2 Required file completeness check: Check whether the unzipped folder contains all required files;

[0073] S2.3 Data size check: Verify that the data file size is within a reasonable range;

[0074] S2.4 File naming correctness check: Verify whether the file naming conforms to the preset naming convention.

[0075] In S3, the storage strategy of separating data files and coordinate information is as follows: the original image data is stored in a distributed file system or object storage, and the coordinate information, metadata and quality inspection results are stored in a spatial database.

[0076] The quality levels are based on administrative divisions and time constraints, ranking data quality with cloud coverage <5% as excellent, <10% as good, and <20% as average.

[0077] S4 specifically includes:

[0078] S4.1: Parse the spatial coverage of the currently accessed data from the stored metadata;

[0079] S4.2: Perform a spatial intersection operation between the spatial coverage area and the preset target area to determine the administrative region to which it belongs; the formula for the spatial intersection operation is:

[0080] Containment relationship: Area(I ∩ C) = Area(I)

[0081] Comparison:

[0082] Where I represents access data, C represents administrative division data, and Area() represents the area calculation function;

[0083] S4.3: Determine whether the data coverage area completely includes or equals the target area, using the following formula:

[0084]

[0085] in, represent , represents the effective area of ​​the nth image, that is, the access data without cloud coverage, and ε represents the allowable uncovered area. For example, if the allowable uncovered area is 1%, then ε is 0.01.

[0086] S4.4: If the data is determined to be fully covered, a processing task will be automatically generated and subsequent data processing will be triggered. If the data is determined to be partially covered, it will be marked as data to be stitched together and will wait to be combined with other data to form full coverage before the processing process is triggered. For partially covered data, the system will automatically search for stored data from adjacent time phases or adjacent orbits for stitching. When the stitched data meets the full coverage requirements, the processing process will be automatically triggered.

[0087] S4.5: Construct an evaluation index system, including six major indicators: cloud coverage ratio, spatial resolution, signal-to-noise ratio, temporal suitability, data source priority, and radiation consistency.

[0088] S4.6: Each image is scored using a comprehensive scoring model.

[0089] Wherein, Score is the overall score, c is the cloud coverage ratio, R is the spatial resolution, SNR is the signal-to-noise ratio, T is the temporal suitability, S is the data source priority, and Rad is the radiation consistency. Assigning weights to each indicator. This is the normalization function for each indicator;

[0090] S4.7: Select the image with the highest overall score as the priority for processing.

[0091] Specifically, the normalization function for each indicator in the multi-indicator comprehensive evaluation model is as follows:

[0092] Cloud coverage ratio ( ):

[0093]

[0094] in, This is a function to normalize the cloud coverage ratio. The maximum allowable cloud cover threshold, such as 20%.

[0095] Spatial resolution ( ):

[0096]

[0097] in, This is the spatial resolution normalization function. and These are the maximum and minimum resolutions of all images in the system, respectively.

[0098] Signal-to-noise ratio (SNR) ):

[0099]

[0100] in, This is the signal-to-noise ratio (SNR) normalization function, where the SNR is estimated as the ratio of the image's mean to its standard deviation. For an ideal signal-to-noise ratio, This represents the minimum acceptable signal-to-noise ratio.

[0101] Timing suitability ( ):

[0102]

[0103] in, For phase suitability normalization function, The optimal time phase (such as the vegetation growing season). For phase tolerance, It is a natural exponential function.

[0104] Data source priority ( ):

[0105]

[0106] in, This is a function for normalizing the priority of the data source. The priority sort value for the current data source. The sorting value corresponding to the highest priority among all candidate data sources is preset by the user, such as: Planet=1, Sentinel-2=0.9, Landsat=0.8.

[0107] Radiation uniformity ( )

[0108]

[0109] in, Let be the radiation uniformity normalization function. This represents the average value of the current image band b. This is the historical average for the same region and frequency band. This refers to the number of bands.

[0110] The system offers three pre-defined standardized weight configuration modes to meet different business needs. Each mode adjusts the weight distribution of six evaluation indicators to precisely tailor the focus of image selection.

[0111] 1. Standard mode: cloud coverage (0.30), spatial resolution (0.25), signal-to-noise ratio (0.15), temporal suitability (0.10), data source priority (0.10), and radiation consistency (0.10). It balances data quality, timeliness, and data source adaptability, and is suitable for most general scenarios.

[0112] 2. Quality Priority Mode: Cloud Coverage (0.35), Signal-to-Noise Ratio (0.25), Radiometric Consistency (0.20), Spatial Resolution (0.10), Temporal Suitability (0.05), Data Source Priority (0.05) This mode focuses on strengthening the weight of basic image quality indicators to ensure that the images input into the S5 processing flow have high definition, low noise, and radiometric stability, making it suitable for scenarios with extremely high product accuracy requirements.

[0113] 3. Timeliness Priority Mode: Timing suitability (0.35), cloud coverage (0.25), data source priority (0.15), spatial resolution (0.10), signal-to-noise ratio (0.10), and radiometric consistency (0.05) prioritize the matching degree between image acquisition timing and business needs. It is suitable for scenarios with urgent timeliness requirements, such as disaster emergency monitoring and short-term dynamic change analysis.

[0114] The system records the images that enter the S5 processing flow after each S4 optimization, as well as the actual quality score of the remote sensing processing product output by S5, and compares it with the preset expected quality score. Based on the deviation between the two, the weights of each evaluation indicator are iteratively updated using the following optimization formula:

[0115]

[0116] in, For the first The updated weight values ​​of the evaluation indicators For the first The original weight values ​​of the evaluation indicators before the update For learning rate, To score the actual product quality, Assuming a quality score, To calculate the overall score for the first The partial derivatives of the weights of the evaluation indicators. Through this mechanism, the weight configuration can be continuously optimized based on the actual operating results of the system, making subsequent image selection more aligned with the needs of business scenarios.

[0117] In S5, the data processing flow specifically includes:

[0118] S5.1: Radiometric calibration, which converts the raw digital quantization values ​​of an image into physically meaningful surface reflectance or radiance values;

[0119] S5.2: Atmospheric correction, eliminating the effects of atmospheric scattering and absorption on remote sensing images;

[0120] S5.3: Geometric correction, eliminating image geometric distortion caused by sensor orientation and terrain undulation;

[0121] S5.4: Image enhancement, which improves image quality through histogram equalization and filtering methods;

[0122] S5.5: Data fusion, which involves fusing multi-source, multi-resolution remote sensing data;

[0123] S5.6: Thematic extraction, using machine learning or deep learning algorithms for land cover classification, change detection, and target recognition.

Claims

1. A method for intelligent access and processing of multi-source remote sensing optical data, characterized in that, include: S1. Automatic Acquisition: Based on the preset task list and scheduling strategy, automatically acquire raw remote sensing optical data from multiple public remote sensing data sources; S2. Data Access and Quality Inspection: Receive the raw remote sensing optical data, perform quality inspection on it, filter out the data that passes the quality inspection and output it. S3. Data storage: Receive the qualified data, store it using a storage strategy that separates data files and coordinate information, and label each data unit with one or more tags including: data source, collection time and quality level. Establish a data index with "administrative division-time" as the basic dimension, sort the data quality, and output the stored structured data with tags and indexes. S4. Processing trigger judgment: Receive the structured data, parse the metadata in it to obtain the spatial coverage of the data, perform spatial intersection calculation with the preset target area administrative division, if the calculation result shows that the access data corresponding to the structured data has achieved full coverage of the target area, then automatically trigger the subsequent data processing process and output the processing task instruction. The spatial intersection operation uses a spatial analysis method based on vector polygons or raster data to calculate the overlap between the data coverage area and the target area. S5. Data Processing: Receive the processing task instruction, call the corresponding qualified data in the structured data in S3, execute the data processing flow, and output the remote sensing processing product.

2. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, In S1, the plurality of publicly disclosed remote sensing data sources include, but are not limited to, at least two of the following: Landsat series, Sentinel series, MODIS, Gaofen series, and Planet data sources; The automatic capture includes at least one of the following: capture period, time range, spatial range, cloud cover threshold, and data resolution.

3. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, S1 specifically includes: S1.1 Automatic data capture initialization, preset capture period, data source list, time range, spatial range, cloud cover threshold parameters, and output preset parameter set; S1.2 Construct and send an API query request, dynamically constructing a query request that conforms to the API interface specification of the target data source based on the preset parameter set; S1.

3. Based on the query request, parse the API response and filter the results to form a list of original data to be downloaded; S1.4 Raw data download: Traverse the list of raw data to be downloaded and initiate a download request to obtain the remote sensing optical raw data; S1.5 Metadata Extraction and Task Log Recording: Extract key information from the downloaded remote sensing optical raw data and generate an execution log.

4. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, S2 specifically includes: S2.

1. Check the integrity of the compressed file to verify whether the decompression process is normal. S2.2 Required file completeness check: Check whether the unzipped folder contains all required files; S2.3 Data size check: Verify that the data file size is within a reasonable range; S2.4 File naming correctness check: Verify whether the file naming conforms to the preset naming convention.

5. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, In S3, the storage strategy of separating data files and coordinate information is as follows: the original image data is stored in a distributed file system or object storage, and the coordinate information, metadata and quality inspection results are stored in a spatial database. The quality levels are based on administrative divisions and time constraints, ranking data quality with cloud coverage <5% as excellent, <10% as good, and <20% as average.

6. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, S4 specifically includes: S4.1: Parse the spatial coverage of the currently accessed data from the stored metadata; S4.2: Perform a spatial intersection operation between the spatial coverage area and the preset target area to determine the administrative region to which it belongs; the formula for the spatial intersection operation is: Containment relationship: Area(I ∩ C) = Area(I); Comparison: ; Where I represents access data, C represents administrative division data, and Area() represents the area calculation function; S4.3: Determine whether the data coverage area completely includes or equals the target area, using the following formula: ； in, represent , represents the effective area of ​​the nth image, that is, the access data without cloud coverage, and ε represents the allowable uncovered area. For example, if the allowable uncovered area is 1%, then ε is 0.

01. S4.4: If the data is determined to be fully covered, a processing task will be automatically generated and subsequent data processing will be triggered. If the data is determined to be partially covered, it will be marked as data to be stitched together and will wait to be combined with other data to form full coverage before the processing process is triggered. For partially covered data, the system will automatically search for stored data from adjacent time phases or adjacent orbits for stitching. When the stitched data meets the full coverage requirements, the processing process will be automatically triggered. S4.5: Construct an evaluation index system, including six major indicators: cloud coverage ratio, spatial resolution, signal-to-noise ratio, temporal suitability, data source priority, and radiation consistency. S4.6: Each image is scored using a comprehensive scoring model. ; Wherein, Score is the overall score, c is the cloud coverage ratio, R is the spatial resolution, SNR is the signal-to-noise ratio, T is the temporal suitability, S is the data source priority, and Rad is the radiation consistency. Assigning weights to each indicator. This is the normalization function for each indicator; S4.7: Select the image with the highest overall score as the priority for processing.

7. The intelligent access and processing method for multi-source remote sensing optical data according to claim 1, characterized in that, In S5, the data processing flow specifically includes: S5.1: Radiometric calibration, which converts the raw digital quantization values ​​of an image into physically meaningful surface reflectance or radiance values; S5.2: Atmospheric correction, eliminating the effects of atmospheric scattering and absorption on remote sensing images; S5.3: Geometric correction, eliminating image geometric distortion caused by sensor orientation and terrain undulation; S5.4: Image enhancement, which improves image quality through histogram equalization and filtering methods; S5.5: Data fusion, which involves fusing multi-source, multi-resolution remote sensing data; S5.6: Thematic extraction, using machine learning or deep learning algorithms for land cover classification, change detection, and target recognition.

8. An electronic device, including a memory and a processor, characterized in that, The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the access and processing method according to any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the access and processing method according to any one of claims 1-7.

Images