Optical high-resolution remote sensing image data screening method and system oriented to global coverage

Abstract

The invention relates to a global coverage-oriented optical high-resolution remote sensing image data screening method and system, and belongs to the technical field of remote sensing image data processing. The method comprises the following steps: dividing a global coverage area and defining a product data set; constructing an original database; executing a first round of quality screening process; original data return is carried out, and warehousing quality inspection and cloud vector area statistical processes are executed; a first round of quality screening process is closed, and warehousing quality inspection and cloud detection are completed; executing a second round of completion screening process, and expanding the time phase constraint; screening and repairing data; after expanding the time phase, screening supplementary data; a second round of completion screening process is closed, and year-round time phase data are combined; executing a third round of optimization screening process, and externally expanding constraint conditions on cloud snow; and the third round of closed-loop optimization screening process supplements cloud vulnerabilities, and supports subsequent low-cloud-cover area one-picture making. According to the invention, unified screening of image data can be realized in a global range, the automation degree is high, and the flexibility is strong.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image data processing, and in particular to a method and system for screening optical high-resolution remote sensing image data with global coverage. Background Art

[0002] Remote sensing satellites, with their unrestricted spatial coverage, wide coverage, short revisit intervals, rapid maneuverable imaging, and high-precision positioning, enable the data and information they capture to play a vital role in the economy, environment, resources, and national security. Satellite optical cameras cannot simultaneously achieve high-resolution imaging, rapid revisits, and observations over larger areas. Therefore, large-scale Earth observation is often achieved through the networking of multiple high-resolution optical remote sensing satellite constellations.

[0003] As the number of satellites in orbit continues to increase, the multi-satellite satellite constellation networking model can achieve high-resolution coverage imaging of the entire global area every year. In order to integrate a set of globally unified advanced product data results, it is necessary to screen out raw data images with single-layer coverage on a global scale from the massive amount of data, and mosaic and splice the image data of adjacent areas. However, due to the influence of various factors such as sensor performance, image phase, lighting changes, and cloud and snow cover, the quality of remote sensing image raw data may vary. In order to obtain global high-resolution optical remote sensing image advanced product results with accurate positioning accuracy and high apparent quality, it is necessary to formulate strict screening standards and optimal single-layer coverage strategies during the raw data screening process to reduce the difficulty of subsequent advanced product production and processing.

[0004] In terms of screening global optical high-resolution remote sensing image data, existing methods have the following shortcomings:

[0005] (1) Image quality varies:

[0006] Existing screening methods typically rely on image quality metrics such as resolution and signal-to-noise ratio to select data, but fail to fully consider the impact of changing conditions such as season, weather, and geographical environment on image quality. For example, cloud cover, snow cover, and lighting differences can adversely affect image quality, resulting in the selected images being unable to meet the requirements of high-precision Earth observation, thereby reducing the quality and consistency of image stitching.

[0007] (2) Poor consistency in data screening:

[0008] Existing screening methods lack the requirement for temporal consistency of images taken within the same region, and thus cannot ensure the consistency of image characteristics such as hue and texture across the same region. This results in noticeable color differences and discontinuities during the image stitching process, affecting the visual quality of the images and the reliability of subsequent analysis.

[0009] (3) Insufficient completeness of coverage:

[0010] Because image acquisition is often limited by satellite capture times and orbital coverage, existing methods struggle to achieve seamless, single-layer coverage across the globe. Data often contains regional overlaps or gaps, making it difficult to generate seamless, single-layer global imagery. This requires additional data supplementation and processing, increasing processing complexity.

[0011] (4) Insufficient optimization of revisit time:

[0012] Current methods fail to adequately consider the optimization of revisit times for high-resolution satellite imagery. While multi-satellite constellations provide frequent observations, the existing selection process struggles to optimize image selection based on revisit times, resulting in a lack of optimal temporal consistency. Particularly in regions with significant seasonal variations, images taken at different times can exhibit significant tonal variations, impacting the overall consistency of the data.

[0013] (5) The screening criteria are not flexible enough:

[0014] Existing screening strategies are often rigid, making it difficult to flexibly adjust screening criteria based on the geographic characteristics of different regions and specific mission requirements. For example, some regions prioritize low cloud cover data, while others prioritize imaging resolution and positioning accuracy. This lack of targeted screening criteria makes the screening process unable to adapt to the unique requirements of different regions and missions.

[0015] (6) The integration of multi-source image data is difficult:

[0016] Existing methods struggle to effectively integrate multi-source imagery captured by multiple satellites. Due to differences in sensor performance, orbital variations, and imaging angles, the images acquired by each satellite are difficult to visually match. Existing methods struggle to overcome these discrepancies when stitching images, often resulting in image discontinuities and blurred edges.

[0017] (7) Lack of automated and intelligent screening support:

[0018] Current screening processes mostly rely on manual labor, resulting in low screening efficiency and difficulty in rapidly processing massive amounts of remote sensing image data. While some methods utilize automated tools, they lack advanced technologies like deep learning to assist with automated image quality analysis and intelligent selection of optimal data. This results in low screening efficiency, making it difficult to meet the demands for efficient and accurate data collection, particularly when dealing with global imagery. Summary of the Invention

[0019] To address the above problems, the present invention proposes a remote sensing image data screening method that can uniformly screen on a global scale, has a high degree of automation, and is highly flexible, so as to solve the current challenges faced in the remote sensing image screening process, such as uneven quality, incomplete coverage, and difficulty in splicing.

[0020] In order to solve the above problems, the present invention adopts the following technical solutions:

[0021] A method for screening optical high-resolution remote sensing image data with global coverage, the method comprising the following steps:

[0022] Step 1: Divide global coverage and define product datasets;

[0023] Step 2: Perform comprehensive image quality assessment on the original data and build an original database;

[0024] Step 3: Perform the first round of quality screening process;

[0025] Step 4: Migrate the original data back, perform the incoming quality inspection and cloud vector area statistics process;

[0026] Step 5: Close the first round of quality screening process and complete incoming quality inspection and cloud testing;

[0027] Step 6: Execute the second round of completion and screening process and expand the time phase constraints;

[0028] Step 7: Screen and repair the data with quality inspection grade B;

[0029] Step 8: Screening and supplementing data after expanding the time phase;

[0030] Step 9: Close the loop of the second round of screening process to merge the full-year temporal data;

[0031] Step 10: Execute the third round of optimization and screening process, and expand the constraints on cloud and snow;

[0032] Step 11: Close the loop of the third round of optimization and screening process to fill in cloud gaps and support the subsequent production of a single map of low cloud cover areas.

[0033] Accordingly, the present invention also proposes a system for screening optical high-resolution remote sensing image data with global coverage, which is used to execute the above-mentioned method to complete the screening of optical high-resolution remote sensing image data with global coverage.

[0034] Compared with the prior art, the present invention has the following beneficial effects: in the production of advanced products in large global regions, the quality of image mosaicking is mainly affected by plane accuracy and apparent color consistency. Therefore, in order to reduce the differences between images and improve the effective use of data, the global data screening method prioritizes overall consistency. That is, based on the characteristics of long satellite photography strips and large image model coverage, the principle of consistent plane positioning accuracy and color of continuous images on the same track is used to ensure same-track coverage. Therefore, the data screening strategy is based on satellite same-track coverage, image quality, cloud and snow, and time phase, and different priorities are screened and counted in the original database. The system has visual selection and customized condition design. After each round of automated screening is completed, manual assistance can be used to check and confirm, and data that does not meet the conditions will be discarded. Based on the above screening principles, the original data is screened multiple times and the database and screening conditions are updated in real time until full coverage is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a flow chart of the method for screening optical high-resolution remote sensing image data with global coverage according to the present invention;

[0036] Figure 2 Schematic diagram of the first round of quality screening process. DETAILED DESCRIPTION

[0037] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments.

[0038] The core of the method for screening optical high-resolution remote sensing image data with global coverage proposed in this paper is based on the image selection process of "first round of quality, second round of completion, and third round of optimization". The specific steps of this method are as follows:

[0039] Step 1: Divide global coverage and define product datasets.

[0040] Step 1.1: To rationally allocate computing resources for massive amounts of data and meet the large-scale production needs of subsequent advanced remote sensing imagery product processing, the global coverage area is divided into regions, and data screening is performed based on survey areas. Survey areas are divided into million-unit units based on national border vectors and topography. The survey area should be as evenly distributed as possible, within 200,000-300,000 square kilometers, and should primarily be regular shapes such as squares or rectangles, while also taking into account the integrity of special terrain such as islands and high mountains.

[0041] Step 1.2: Data is screened based on the regional vector and shooting phase requirements. Without distinguishing the quality level, all products in the specified time interval (e.g., 2023, 2024) are stored in the database to form database S1.

[0042] Step 1.3: Set the temporal database T (including T1, T2, and T3). For example, for the geographical range above 30 degrees north latitude, set the temporal database for June to September as T1, and store the images from June to September in database T1; for the geographical range above 30 degrees south latitude, set the temporal database for November, December, January, and February as T2, and store the images from November, December, January, and February in database T2; for other areas, set the temporal database as T3, and store the images from other months in database T.

[0043] Step 1.4: Generate a survey area vector set B, where the element Bi in the survey area vector set B represents the i-th survey area in the set B.

[0044] Step 2: Set the original data single-view image quality rules and the original database.

[0045] Step 2.1: Perform comprehensive image quality assessment on the panchromatic and multispectral images of the original data based on factors affecting image quality. The panchromatic image quality is affected by the number loss level, RPC convergence level, uncontrolled positioning accuracy level, and color difference level. The multispectral image quality is affected by the number loss level, RPC convergence level, uncontrolled positioning accuracy level, pseudo color level, and color difference level.

[0046] Based on the aforementioned factors affecting image quality, image quality parameters are graded into four levels: A, B, C, and D. A is considered optimal, A and B are acceptable, and C and D are unacceptable. The overall image quality assessment rule is based on the "wooden barrel" principle, with the worst value of each grade being used as the overall image quality.

[0047] From the massive amount of raw data images, select the data with comprehensive image quality of A or B, which meets the requirements of phase, sway, resolution and regional vector range throughout the year, and store them in the raw database;

[0048] Step 2.2: Select a measurement area Bi from the measurement area vector set B. Based on the temporal database T and according to the screening requirements, select all products that intersect with the measurement area Bi and meet the preset conditions (such as cloud cover (≤10%) and side swing ([-15, 15])). Store the product information in database S2.

[0049] Step 3: Perform the first round of quality screening process.

[0050] Step 3.1: In the "First Round Quality" screening, the overall image quality, side swing requirements, resolution, and regional vector range settings remain unchanged, and strict control is exercised over the time phase and cloud and snow conditions to select the best imagery. The cloud cover requirement for the first round is 0. There are areas of perennial snow cover worldwide, and to achieve rich imagery, clear textures, and uniform tones, the best color effects are achieved when the time phases are close and during the vegetation growing season. Therefore, for geographic areas above 30 degrees north latitude, the time phase is set to June-September; for geographic areas above 30 degrees south latitude, the time phase is set to November, December, January, and February; and for other areas near the equator, the time phase is set to year-round.

[0051] Since the results that meet the above requirements have multiple layers of coverage, it is necessary to select the optimal single-layer coverage image. At the same time, in order to meet the subsequent mosaic operation, there should be no less than 4% overlap between adjacent scene images.

[0052] The global data screening method prioritizes overall consistency and uses co-orbital data whenever possible to reduce later misalignment and apparent discrepancies. First, data with the largest total coverage in the same orbit is prioritized; second, data with similar temporal phases is prioritized. Finally, satellite sources with large swaths are prioritized. Using the imaging track as the unit, the area falling within the vector range of the screening region is counted and sorted from large to small. The co-orbital image corresponding to the maximum area coverage is used as the basic data source for coverage, and the remaining area is expanded by a certain number of pixels as the gap area. The sorting and screening, and gap area updates are continuously performed according to the above method until the optimal single-layer coverage is achieved.

[0053] Based on the above principles, the image data with the best single-layer coverage based on the same-track images can be selected. However, problems such as cloud and snow masks, inter-scene colors, and thin clouds and mist cannot be accurately screened through automated strategies. Therefore, the selected images are visualized, and unqualified data are manually discarded to update the vulnerability areas.

[0054] Step 3.2: Based on the database S2, taking a single imaging track as a unit (a wide-area small scene is considered as a single-track imaging task), count the areas falling within the vector range of the screening region, and obtain the sorted set V1 in descending order.

[0055] Step 3.3: Take the task track number corresponding to the maximum value of the sorted set V1 as the basic reference data source for the selection, recorded as basic data B1;

[0056] Step 3.4: Perform a negative buffer operation on the basic data B1, shrinking it by 2 kilometers. Combine the mission vector area and the vector after the reduction of the basic data B1 to form the vulnerability vector B2 as the screening mission area vector. Repeat steps 3.1 to 3.4 to gradually complete the mission area vector data screening and record it in database S3.

[0057] Step 4: Migrate original data back.

[0058] Carry out data migration on database S3, perform incoming quality inspection and cloud vector area statistical processing, eliminate data with overall quality inspection level lower than incoming quality, record it as database SF (feedback SF to the production quality department and L1 production operation department), and form database S4 (including effective coverage area, cloud vector and loophole area).

[0059] Step 5: Closed-Loop "First-Round Quality" Process

[0060] Step 5.1: Based on database S4, continue to search for all scenes based on database S2 and perform quality inspection process. Repeat steps 3 and 4 to complete the "first round" of quality data screening, statistical information of all scenes, scenes that meet the requirements, scenes that are eliminated, and scenes that are stored, and obtain the vulnerability and cloud vector database S5;

[0061] Step 5.2: After the database SF is fed back to the L1 production operations department, the L1 production operations department will carry out data repair work, which is required to take less than 3 days. After receiving the repair completion notice and archiving, the department will re-perform the incoming quality inspection and cloud testing to form the database S6.

[0062] Step 6: Execute the second round of completion and screening process and expand the time constraints.

[0063] The second round of supplementary screening is used to filter and supplement data in areas with gaps. The presence of clouds and snow can cause loss of ground feature information. Based on the priority of large-area image mosaic quality, this round opts for an expansion of the temporal phase. Above 30 degrees north latitude, the temporal phase is set to May–October; above 30 degrees south latitude, the temporal phase is set to October, November, December, January, February, and March. Other screening principles remain the same as the first round of quality screening. After the automated screening is complete, the selected second-round image results are visualized, and unqualified data is manually discarded to update the gap areas.

[0064] Step 7: Start the product repair process.

[0065] Step 7.1: Combine database S6 and database T to conduct a screening process for data with a quality inspection level of B. Other screening restrictions are the same as in step 1. After screening, database S7 is formed. Then, manual data repair is carried out to repair the data with a quality inspection level of B to a quality inspection level of A.

[0066] Step 7.2: After receiving the restoration completion notification and archiving, re-carry out the incoming quality inspection and cloud inspection, eliminate the data with the overall quality inspection grade lower than the incoming quality, record it in database SF2, and form database S8;

[0067] Step 8: Expand product phase and screen supplementary data.

[0068] Step 8.1: Expand the time phase by one month in both directions and filter only the data of the expanded time phase. The corresponding database is recorded as ES1. The screening principle is the same as the process of quality inspection level A. Based on database S8, database S9 and database SF3 are formed;

[0069] Step 8.2: Feedback the database SF3 to the production quality department and the L1 production operation department. Due to time constraints, the advanced production operation will not wait for the repair results.

[0070] Step 9: Close the second round of screening process and merge the full-year temporal data.

[0071] Step 9.1: Merge database S1 with database ES1, change the quality inspection level to B, and repeat steps 2 and 3 to form vector database S10;

[0072] Step 9.2: Use data with a quality level of A+B for the entire year to fill in the gaps and perform quality inspection on the incoming data. At the same time, record the data with L1 quality level that does not meet the incoming quality inspection as vector database SF4, and update the data to cover the vector database S11 (full time phase, A+B optimal data).

[0073] Step 10: Execute the third round of optimization and screening process, and expand the constraints on cloud and snow.

[0074] Step 10.1: The third round of screening filters and supplements data in the vulnerability area. This round expands the cloud and snow constraints. Other screening principles remain the same as the second round of complete screening process. The cloud and snow content can be determined based on the task requirements, generally set to less than 10%. After the automated screening is complete, the selected third round image results are visualized. Unqualified data is manually discarded, and a combined cloud and snow vector is generated based on the cloud and snow detection algorithm. The unqualified data vector and the cloud and snow vector are processed through the connected domain to form the vulnerability. Repeat this screening process for the vulnerability area until the entire task area is fully covered.

[0075] Step 10.2: To ensure consistent edge coverage between survey intervals, adjacent survey intervals must be connected on the same track. During data screening, prioritize the upper left and lower right directions for two-track edge coverage. This ensures full coverage of adjacent survey intervals.

[0076] Step 11: Close the loop of the third round of optimization and screening process to achieve low cloud coverage across the entire region.

[0077] Step 11.1: Set the cloud cover index to less than 10%. When there are gaps in database S11, use data with a quality lower than level B to supplement it, forming a data vector and database S12 that fully covers the mission area.

[0078] Step 11.2: Merge the cloud vectors in database S12 and perform connected domain processing to form cloud vulnerability area S13. Based on database S2, perform incoming quality inspection according to cloud vulnerability area S13, supplement cloud vulnerabilities, and finally form database S14 to support the subsequent production of a map of low cloud cover areas.

[0079] The present invention also proposes a system for screening optical high-resolution remote sensing image data with global coverage, which is used to execute the various steps recorded in the method for screening optical high-resolution remote sensing image data with global coverage as described above to complete the screening of optical high-resolution remote sensing image data with global coverage.

[0080] The optical high-resolution remote sensing image data screening method and system for global coverage proposed in the present invention are designed to solve the problems of wide geographical scope, complex topography, and large temporal span of the global region. Based on the principle of overall consistency priority, the same-track images are sorted by the total coverage area of ​​the region, and priority is given to large-area same-track images with high comprehensive image quality evaluation for coverage. This method has the characteristics of low computational complexity and high degree of automation. At the same time, the same-track priority can ensure the subsequent geometric positioning accuracy and the quality of apparent color difference consistency and reduce the complexity of mosaic splicing. The present invention can directly screen data of massive global original images, and formulates a screening strategy for optimal coverage of high-quality data by setting indicators such as temporal phase, cloud and snow, and original image quality. The system screening based on this method has a high degree of automation and high efficiency. Combined with manual inspection, it can judge and analyze areas in different geographical locations and select the optimal data for screening.

[0081] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be determined by the appended claims.

Claims

1. A method for screening optical high-resolution remote sensing image data with global coverage, characterized in that: The following steps are involved: Step 1: Divide global coverage and define product datasets; Step 2: Perform comprehensive image quality assessment on the original data and build an original database; Step 3: Perform the first round of quality screening process; Step 4: Migrate the original data back, perform the incoming quality inspection and cloud vector area statistics process; Step 5: Close the first round of quality screening process and complete incoming quality inspection and cloud testing; Step 6: Execute the second round of completion and screening process and expand the time phase constraints; Step 7: Screen and repair the data with quality inspection grade B; Step 8: Screening and supplementing data after expanding the time phase; Step 9: Close the loop of the second round of screening process to merge the full-year temporal data; Step 10: Execute the third round of optimization and screening process, and expand the constraints on cloud and snow; Step 11: Close the loop of the third round of optimization and screening process to fill in cloud gaps and support the subsequent production of a single map of low cloud cover areas.

2. The method for screening global optical high-resolution remote sensing image data according to claim 1, characterized in that: Step 1 includes the following steps: Step 1.1: Divide the global coverage area into regions and perform data screening based on the survey area. The survey area is divided into millions of sections based on national border vectors and topography. Step 1.2: Based on the regional vector and shooting phase requirements, without distinguishing the quality levels, all products in the specified time interval are stored in the database to form database S1; Step 1.3: Set the temporal database T. For the geographical range above 30 degrees north latitude, set the temporal database for June to September as T1; for the geographical range above 30 degrees south latitude, set the temporal database for November, December, January, and February as T2; the temporal database for other regions is recorded as T3. Step 1.4: Generate the survey area vector set B.

3. The method for screening global optical high-resolution remote sensing image data according to claim 2, characterized in that: Step 2 includes the following steps: Step 2.1: Assess the image quality parameters based on the factors affecting image quality, and complete the comprehensive image quality assessment of the original panchromatic and multispectral images. From the massive amount of raw data images, select the data with an overall image quality of A or B and that meet the requirements of phase, side swing, resolution, and regional vector range throughout the year, and store them in the raw database; Step 2.2: Select one of the survey areas Bi in the survey area vector set B. Based on the temporal database T and according to the screening requirements, select all products that intersect with the survey area Bi and meet the preset conditions, and store the product information in the database S2.

4. The method for screening global optical high-resolution remote sensing image data according to claim 3, characterized in that: Step 3 includes the following steps: Step 3.1: Maintaining the same image quality, side swing requirements, resolution, and regional vector range settings, strictly control the temporal phase and cloud and snow conditions, select the image data with the best single-layer coverage based on the same-track image, manually discard unqualified data, and update the vulnerability area; Step 3.2: Based on the database S2, take a single imaging track as a unit, count the areas that fall within the vector range of the screening region, and obtain the sorted set V1 in descending order; Step 3.3: Take the task track number corresponding to the maximum value of the sorted set V1 as the basic reference data source for the selection, recorded as basic data B1; Step 3.4: Perform a negative buffer operation on the basic data B1, shrinking it by 2 kilometers. Combine the mission vector area and the vector after the reduction of the basic data B1 to form the vulnerability vector B2 as the screening mission area vector. Repeat steps 3.1 to 3.4 to gradually complete the mission area vector data screening and record it in database S3.

5. The method for screening global optical high-resolution remote sensing image data according to claim 4, characterized in that: Step 4 includes: Data is migrated back to database S3, and the quality inspection and cloud vector area statistical processing process are performed. Data with the whole scene quality inspection level lower than the quality of the entry are eliminated and recorded as database SF to form database S4.

6. The method for screening global optical high-resolution remote sensing image data according to claim 5, characterized in that: Step 5 includes the following steps: Step 5.1: Based on database S4, continue to search for all scenes based on database S2 and perform quality inspection. Repeat steps 3 and 4 to complete the "first round" of quality data screening. Collect statistics for all scenes, scenes that meet the requirements, scenes that are rejected, and scenes that are stored. Finally, obtain the vulnerability and cloud vector database S5. Step 5.2: After SF is fed back to the L1 production operation department, the L1 production operation department will carry out data repair processing. After receiving the repair completion notification and archiving, it will re-carry out warehousing quality inspection and cloud detection to form database S6.

7. The method for screening global optical high-resolution remote sensing image data according to claim 1 or 2, characterized in that: Step 6 includes the following steps: The second round of screening screens and supplements data in the vulnerability area and expands the time phase. Other screening principles are the same as the first round of quality screening conditions. After the automated screening is completed, the selected second round image results are visualized, and unqualified data are manually discarded to update the vulnerability area.

8. The method for screening global optical high-resolution remote sensing image data according to claim 6, characterized in that: Step 7 includes the following steps: Step 7.1: Combine database S6 and database T to conduct a screening process for data with a quality inspection level of B. Other screening restrictions are the same as in step 1. After screening, database S7 is formed. Then, manual data repair is carried out to repair the data with a quality inspection level of B to a quality inspection level of A. Step 7.2: After receiving the restoration completion notification and archiving, re-conduct the incoming quality inspection and cloud detection, and eliminate the data with the overall quality inspection grade lower than the incoming quality and record it in database SF2, forming database S8.

9. The method for screening global optical high-resolution remote sensing image data according to claim 1 or 2, characterized in that: Step 10 includes the following steps: Step 10.1: Screen and supplement data in the vulnerability area and expand the cloud and snow constraints. Other screening principles remain the same as the second round of supplementary screening process. After the automated screening is completed, visualize the selected third round image results. Manually discard unqualified data, and obtain a merged cloud and snow vector based on the cloud and snow detection algorithm. After the unqualified data vector and the cloud and snow vector are processed through the connected domain, the vulnerability is formed. Repeat this step of screening for the vulnerability area until the entire task area is covered. Step 10.2: During data screening, the survey area prioritizes the expansion of the two-track same-scene edge data in the upper left and lower right directions, so that all surrounding survey areas can complete the full coverage of the same-track connection.

10. A system for screening optical high-resolution remote sensing image data with global coverage, characterized in that: Used to execute the method according to any one of claims 1 to 9 to complete the screening of optical high-resolution remote sensing image data with global coverage.

Images