Domestic land exploration data automatic framing method based on space geometry overlapping relation

Abstract

The invention discloses a domestic land exploration data automatic framing method based on a space geometry overlapping relation, and the method comprises the steps: obtaining SAR image data which comprises a standard sheet data set and a to-be-framed sheet data set; performing framing on each single-scene standard mapsheet in the labeled mapsheet data set to obtain a plurality of single standard mapsheets; each single standard mapsheet is retrieved in a to-be-framed mapsheet data set, screening is carried out according to a spatial geometric overlapping relation, and images with the spatial overlapping relation are selected as a preliminary screening result; according to the method, on the basis of a standard sheet data set and a to-be-framed sheet data set, a single standard sheet is firstly retrieved in the to-be-framed sheet data set according to a geometric overlapping relation, and then the to-be-framed data set is retrieved according to the relation of the longitude and latitude of angular points; and finally, retrieval is carried out according to the imaging time, the geometric overlapping relation and the overlapping area proportion, so that the utilization rate of the domestic land exploration data is effectively improved.

Description

Technical Field

[0001] This invention relates to the fields of wide-area InSAR and domestic land exploration data tiling processing technology, and particularly to an automated domestic land exploration data tiling method based on spatial geometric overlap. Background Technology

[0002] Satellite remote sensing technology, as an advanced means of information acquisition, plays an increasingly important role in modern human production and life, involving multiple fields and bringing significant social and economic benefits, providing information support for human production, life, and ecological space construction. Against the backdrop of the national "Aerospace Information Construction" initiative, the demand for using satellite remote sensing technology for urban building safety risk assessment, digital earth, and large-scale, wide-area temporal deformation monitoring has rapidly emerged, creating a market demand worth hundreds of billions of yuan. However, due to the large monitoring area and the complex and diverse types of regional landforms, it is difficult to conduct large-scale risk assessments manually. With the continuous maturation and development of satellite remote sensing technology, combined satellite remote sensing technology has become a new option for large-scale digital earth construction.

[0003] Interferometric Synthetic Aperture Radar (InSAR) technology boasts advantages such as wide coverage, high precision, and high resolution, gradually becoming an effective tool for large-scale air-to-ground observation. Furthermore, InSAR technology can monitor minute surface deformations with high precision, such as natural phenomena like earthquakes, subsidence, and landslides. This capability provides crucial data support for early warning and risk assessment of geological hazards, helping to reduce potential losses. In urban and infrastructure management, regularly acquired InSAR data enables real-time monitoring of building and infrastructure deformation, timely detection of safety hazards, and support for maintenance and reinforcement efforts.

[0004] In SAR image processing, time-series SAR imagery refers to continuous image data of the same area acquired at different times by a synthetic aperture radar (SAR) sensor. Processing long-term time-series SAR image data can reflect the changing surface features and dynamic information of the target area over time. However, not all time-series SAR images are suitable for time-series processing. As satellite monitoring time increases and the amount of data grows, extracting SAR image data suitable for InSAR processing from massive datasets is crucial.

[0005] Past research has typically focused on SAR image content retrieval, specifically identifying SAR images containing specific targets and classifying them into categories. These studies have included using fly-by-flight algorithms to generate hash codes for targets in images, enabling retrieval of land cover classification images in public datasets like OpenSAR and MSTAR. Additionally, researchers have explored retrieval based on deep learning convolutional neural networks (CNNs) and SAR image coherence. However, these studies have primarily concentrated on SAR image content classification, with little focus on related issues in the temporal InSAR processing.

[0006] The latest retrieval algorithm, focusing on temporal InSAR data processing, is designed for foreign sentinel imagery data. Based on a traditional greedy algorithm, it retrieves the optimal map coverage for the Region of Interest (ROI) to reduce data redundancy during processing. For land exploration data, due to orbital control factors (time-based image segmentation) and SAR sensor position offsets, temporal SAR images of the same area exhibit spatial discontinuities. Existing methods cannot automate the retrieval and classification of these discontinuous SAR images; currently, classification and retrieval are mostly performed manually. When faced with large-scale, wide-area InSAR processing tasks, manual retrieval and classification methods are time-consuming, labor-intensive, and extremely inefficient. To automate and rapidly perform wide-area InSAR processing of land exploration data, it is necessary to analyze the imaging positions of various satellite images and retrieve the required SAR imagery data based on appropriate geometric overlap relationships. Then, the retrieved images are stitched and cropped to maximize the use of SAR data and achieve wide-area InSAR data processing. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide an automated method for processing domestic land exploration data based on spatial geometric overlap. By utilizing the spatial geometric overlap between map sheets, combined with imaging time and coverage ratio, the method can search and filter massive amounts of SAR data, effectively improving the utilization rate of domestic land exploration data.

[0008] To achieve the above objectives, the technical solution of the present invention is as follows:

[0009] An automatic zoning method for domestic land exploration data based on spatial geometric overlap includes:

[0010] Acquire SAR image data, which includes a standard map sheet dataset and a map sheet dataset to be divided into sheets;

[0011] Each scene standard map sheet in the labeled map sheet dataset is divided into several individual standard map sheets;

[0012] Each of the individual standard map sheets is retrieved from the dataset of map sheets to be divided, and filtered according to spatial geometric overlap. Images with spatial overlap are selected as preliminary filtering results. The preliminary filtering results include the track data of the left and right sides of the standard map sheet.

[0013] A corner longitude threshold is set on the preliminary screening results to remove map data on other tracks besides the left and right sides, thus obtaining the first map dataset;

[0014] A corner latitude threshold is set for the first map sheet dataset to remove redundant map sheet datasets from the upper and lower standard map sheets, thus obtaining the second map sheet dataset.

[0015] The second map dataset is filtered using time, spatial geometric overlap, and overlap area ratio to obtain the split map dataset.

[0016] Optionally, the filtering of the second map dataset based on temporal and spatial geometric overlap includes:

[0017] Sort the second map dataset in chronological order and find two map images with the same date and time.

[0018] Determine whether two image frames with the same date and time have overlapping areas. If they have overlapping areas, the two image frames form a time-series SAR image.

[0019] Determine whether the overlap ratio between the time-series SAR image and the standard map sheet is greater than a set threshold; if it is, retain it.

[0020] Optionally, the step of filtering the second map dataset using the overlap area ratio includes:

[0021] For each single-scene image in the second map sheet dataset, an analysis of the spatial overlap area ratio with the standard map sheet is performed. Data whose spatial geometric overlap area does not reach the set area ratio is deleted, while data that overlaps extensively with the standard map sheet is retained.

[0022] Optionally, the SAR image data is sourced from the L-band Land Observation Satellite-1.

[0023] Optionally, the labeled map dataset refers to the map locations covering the study area.

[0024] Optionally, the dataset of map sheets to be divided includes image data covering all study areas within the imaging time frame.

[0025] Optionally, the SAR image data can be acquired in the following manner:

[0026] Based on the requirements, first select the coverage area of ​​the region of interest in the Land Observation Satellite Service Center, including file input, administrative region selection, and custom polygon selection;

[0027] Select the satellite imaging mode, including parameter settings such as mode selection, land observation satellite sensor settings, orbit type, and setting the time range for image acquisition;

[0028] Download the selected product number and original data information from the Land Observation Satellite Service Center.

[0029] Optionally, the method further includes:

[0030] The segmented map dataset is cropped and stitched together to obtain time-series SAR image data on the standard map sheet.

[0031] Optionally, the method further includes importing the obtained time-series SAR image sheet dataset and standard image sheet data into the Land Observation Satellite Service Center website for verification.

[0032] Optionally, the threshold is 90%.

[0033] Compared with the prior art, the advantages of this invention are as follows:

[0034] This method is based on a standard map sheet dataset and a dataset of maps to be divided. First, a single standard map sheet is searched in the dataset of maps to be divided according to geometric overlap. Then, the dataset of maps to be divided is searched according to the latitude and longitude of the corner points. Finally, for special cases, the search is performed according to imaging time, geometric overlap, and overlap area ratio to obtain the final map division result, which effectively improves the utilization rate of domestic land exploration data. Attached Figure Description

[0035] Figure 1 The main flowchart of the automated domestic land exploration data slicing processing method based on spatial geometric overlap relationship provided in the embodiments of this application;

[0036] Figure 2 A technical roadmap for an automated domestic land exploration data slicing processing method based on spatial geometric overlap provided in this application embodiment;

[0037] Figure 3-7 Spatial location map of various images to be divided into standard map sheets;

[0038] Figure 8 This is a standard map showing the distribution of Yunnan Province.

[0039] Figure 9 This is a standard map showing the distribution of map sheets in Guizhou Province.

[0040] Figure 10 This is a standard map showing the distribution of map sheets in Sichuan Province.

[0041] Figure 11-12 This is a map showing the map cover relationship after a standard map sheet has been filtered by geometric overlap, corner points, and special cases. Detailed Implementation

[0042] Example:

[0043] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0044] For decades, most of the interferometric satellites used in my country were foreign (e.g., Sentinel-1, TerraSAR data), and my country did not yet have a single remote sensing SAR satellite capable of independent interferometric measurements. However, the localization of remote sensing data is of great significance to my country's future national security, independent innovation capabilities, and economic development. With the continuous development of my country's science and technology, my country's first L-band satellite with independent interferometric capabilities, Land Survey Satellite-1 (LT-1), was launched. LT-1 is a remote sensing satellite independently developed by my country with independent interferometric measurement capabilities, mainly used for Earth observation and resource monitoring. Its design and launch aimed to enhance the country's technological capabilities and application level in the field of remote sensing. The satellite was successfully launched in 2022, including a dual-satellite formation flight of LT-1A and LT-1B, marking further progress in my country's "Earth observation" technology. LT-1's technical parameters are quite advanced, possessing high spatial resolution and the ability to clearly capture ground images. This makes it widely applicable in various fields such as urban planning, agricultural monitoring, and environmental protection. The satellite is equipped with a multispectral sensor covering the visible and near-infrared bands, enabling it to acquire rich spectral information and support diverse application needs.

[0045] As more and more SAR data accumulates over time, the inventors have focused on how to efficiently utilize this SAR satellite remote sensing data to generate economic and application effects on society.

[0046] To address this issue, this application proposes an automated method for processing domestic land exploration data based on spatial geometric overlap. By utilizing the spatial geometric overlap between map sheets, combined with imaging time and coverage ratio, massive amounts of SAR data are retrieved and filtered. This invention is the first to solve the map segmentation problem in the InSAR processing of domestic data, significantly improving the retrieval rate of InSAR time-series processing of domestic data.

[0047] For details, please refer to Figure 1As shown, the automated domestic land exploration data tiling processing method based on spatial geometric overlap mainly includes the following steps:

[0048] 110. Acquire SAR image data, which includes a standard map sheet dataset and a map sheet dataset to be divided into sheets;

[0049] In practice, SAR image data is acquired in the following way:

[0050] Based on the requirements, the first step is to select the coverage area of ​​the region of interest in the Land Observation Satellite Service Center. This includes file input (SHP file of the region of interest), administrative region selection (national, provincial, municipal), and custom polygon selection. Additionally, the satellite imaging mode needs to be selected, with parameter settings including mode selection (STRIP1 mode, STRIP2 mode, etc.), Land Observation Satellite sensor settings (formation mode, Satellite A, Satellite B, etc.), orbit type (ascending orbit data, descending orbit data), and the time range for image acquisition.

[0051] The mode selection is set to STRIP1, satellite sensor A is selected, orbit is set to ascending orbit data, and the time range is set to more than one year. Furthermore, the labeled map sheet dataset refers to the map sheet locations covering the study area, ensuring effective coverage of the entire study area. The map sheet dataset to be segmented comprises all image data covering the study area within the imaging time range.

[0052] 120. Divide each scene standard map sheet in the labeled map sheet dataset into several individual standard map sheets.

[0053] In this way, a single standard map sheet can be obtained, which can then be used as a spatial retrieval standard unit in subsequent steps to search for long-term SAR image data that satisfy the spatial geometric overlap relationship in the entire image dataset to be divided into sheets, thereby obtaining a time-series SAR image sheet dataset of all standard map sheets.

[0054] For domestically produced land exploration data, when the LuTan-1 satellite passes through the coverage area of ​​the standard map sheet, the location of the images taken at certain times does not almost overlap with the standard map sheet. These images partially overlap with the coverage area of ​​the standard map sheet, and some do not overlap at all. This necessitates analyzing whether the images overlap with the standard images during the satellite's revisit cycle, and whether time-series SAR images can be restored by stitching and cropping two overlapping images. In addition, it is also necessary to analyze the spatial geometric overlap relationship between the revisited map sheet and the standard map sheet. The spatial positional relationship is shown in the attached figure. Figure 2 As shown in the figure, the following step 130 is required for this purpose.

[0055] 130. Each of the individual standard map sheets is searched in the dataset of map sheets to be divided, and filtered according to the spatial geometric overlap relationship. Images with spatial overlap relationship are selected as the preliminary screening results. The preliminary screening results include the track data on the left and right sides of the standard map sheet.

[0056] In this step of the map retrieval process, based on the overlap relationship between the standard map sheet and the map sheet to be divided, the map sheets that overlap in spatial location are initially screened out, and the images with spatial overlap are selected as the preliminary screening results.

[0057] Since the initial screening results include track data from both sides of the standard map sheet, and the longitude of the track data on both sides differs significantly from that of the standard map sheet, it is necessary to set a longitude threshold for the corner points in the initial screening results to remove map data from other tracks on both sides. This requires the following step 140.

[0058] 140. Set a corner longitude threshold for the preliminary screening results to remove map data on other tracks besides the left and right sides, and obtain the first map dataset, which is the dataset after being filtered by the track longitude threshold.

[0059] Since the data overlapping with the single-scene standard map sheet data includes map sheet data with significant latitude differences, and these map sheet data belong to the upper and lower scenes of the current standard map sheet, it is necessary to filter the dataset after filtering by orbital longitude threshold according to the latitude interval of the corner points to determine the dataset belonging to the current standard map sheet. This requires the following step 150.

[0060] 150. Set a corner latitude threshold for the first map sheet dataset to remove redundant map sheet datasets that overlap between the upper and lower standard map sheets, and obtain the second map sheet dataset, which is the map sheet dataset after being filtered by geometric overlap relationship and latitude and longitude threshold.

[0061] For the map sheet dataset after being segmented based on geometric overlap and latitude / longitude thresholds, although these map sheets overlap with the standard map sheet, they still require filtering due to spatial discontinuity, the need for stitching and cropping of map sheet data from adjacent time periods, and the need to determine whether individual map sheet images overlap with most areas of the standard map sheet. These various situations can be categorized as follows:

[0062] Scenario 1: The foreground and background images overlap with the standard image frame but do not cover any area;

[0063] Scenario 2: The revisited image only partially / partially covers the standard map sheet;

[0064] Scenario 3: The two scenes overlapped on a standard map sheet;

[0065] Scenario 4: The images of the foreground and background do not completely overlap in the standard map sheet;

[0066] Scenario 5: A single SAR image can largely overlap / completely overlap.

[0067] In response to the various situations mentioned above, the following steps 160 are required.

[0068] 160. The second map dataset is filtered using time, spatial geometric overlap, and overlap area ratio, and the resulting split map dataset is output.

[0069] In this step, by utilizing the temporal and spatial geometric overlap relationship, case one can be eliminated, and cases three and four can be selected. Furthermore, by filtering based on the overlap area ratio, case two can be eliminated, and case five can be selected. The overall technical roadmap of this method is as follows: Figure 2 As shown.

[0070] In summary, this method, based on the standard map sheet dataset and the map sheet dataset to be divided, first searches the map sheet dataset to be divided according to the geometric overlap relationship of a single standard map sheet, then searches the map sheet dataset to be divided according to the latitude and longitude relationship of the corner points, and finally searches according to imaging time, geometric overlap relationship and overlap area ratio for special cases to obtain the final map sheet division result, which effectively improves the utilization rate of domestic land exploration data.

[0071] In a preferred embodiment, the automated domestic land exploration data tiling processing method based on spatial geometric overlap provided in this embodiment further includes:

[0072] The divided map dataset is cropped and stitched together to obtain time-series SAR image data on the standard map sheet;

[0073] The obtained temporal SAR image dataset and standard image data were imported into the Land Observation Satellite Service Center website for verification.

[0074] In the specific implementation process, the filtering of the second map sheet dataset using temporal and spatial geometric overlap relationships includes:

[0075] Sort the second map dataset in chronological order and find two map images with the same date and time.

[0076] To determine whether two image sheets with the same date and time overlap, if they do not overlap, these two image sheets are spatially separated from the standard image sheet, and the two image sheets cannot be stitched or cropped to form a time-series SAR image of the standard image sheet. Figure 3-4As shown. When there are overlapping areas, these two image sheets can be stitched together and cropped to form a standard image sheet time-series SAR image data, such as... Figure 5-6 As shown.

[0077] For time-series SAR imagery data with overlapping areas that are stitched and cropped to form a standard map sheet, the overlap ratio between the stitched and cropped time-series SAR imagery and the standard map sheet is determined. The ratio needs to be set to ensure a large overlap (set to 90%). In this case, such as... Figure 7 As shown. Filtering is based on the overlap area ratio.

[0078] In a specific implementation, the filtering of the second map sheet dataset using the overlap area ratio includes:

[0079] For each single-scene image in the second map sheet dataset, an analysis of the spatial overlap area ratio with the standard map sheet is performed. Data whose spatial geometric overlap area does not reach the set area ratio is deleted, while data that overlaps extensively with the standard map sheet is retained.

[0080] like Figure 8-10 As shown, this embodiment takes the three southwestern provinces of my country (Yunnan, Guizhou and Sichuan) as an example. It utilizes standard map sheet data covering the entire territory of the three provinces, as well as the orbital data of the LT-1AB satellite obtained by the land exploration satellite from July 2023 to September 2024, to realize the segmentation of time-series SAR images of Yunnan, Guizhou and Sichuan provinces. Figure 11-12 The results are shown under a standard map sheet of Sichuan Province, after being filtered by geometric overlap relationship (Figure), latitude and longitude corner point (Figure), and special case (Figure).

[0081] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An automatic zoning method for domestic land exploration data based on spatial geometric overlap, characterized in that, include: Acquire SAR image data, which includes a standard map sheet dataset and a map sheet dataset to be divided into sheets; Each scene standard map sheet in the labeled map sheet dataset is divided into several individual standard map sheets; Each of the individual standard map sheets is retrieved from the dataset of map sheets to be divided, and filtered according to spatial geometric overlap. Images with spatial overlap are selected as preliminary filtering results. The preliminary filtering results include the track data of the left and right sides of the standard map sheet. A corner longitude threshold is set on the preliminary screening results to remove map data on other tracks besides the left and right sides, thus obtaining the first map dataset; A corner latitude threshold is set for the first map sheet dataset to remove redundant map sheet datasets from the upper and lower standard map sheets, thus obtaining the second map sheet dataset. The second map dataset is filtered using time, spatial geometric overlap, and overlap area ratio to obtain the split map dataset.

2. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1, characterized in that, The filtering of the second map dataset using temporal and spatial geometric overlap relationships includes: Sort the second map dataset in chronological order and find two map images with the same date and time. Determine whether two image frames with the same date and time have overlapping areas. If they have overlapping areas, the two image frames form a time-series SAR image. Determine whether the overlap ratio between the time-series SAR image and the standard map sheet is greater than a set threshold; if it is, retain it.

3. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1, characterized in that, The filtering of the second map dataset using the overlap area ratio includes: For each single-scene image in the second map sheet dataset, an analysis of the spatial overlap area ratio with the standard map sheet is performed. Data whose spatial geometric overlap area does not reach the set area ratio is deleted, while data that overlaps extensively with the standard map sheet is retained.

4. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1, characterized in that, The SAR image data was obtained from the L-band satellite Land Exploration Satellite 1.

5. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1, characterized in that, The labeled map dataset refers to the map locations covering the study area.

6. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1 or 5, characterized in that, The dataset of images to be divided contains image data covering all study areas within the imaging time frame.

7. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 4, characterized in that, The SAR image data is obtained in the following manner: Based on the requirements, first select the coverage area of ​​the region of interest in the Land Observation Satellite Service Center, including file input, administrative region selection, and custom polygon selection; Select the satellite imaging mode, including parameter settings such as mode selection, land observation satellite sensor settings, orbit type, and setting the time range for image acquisition; Download the selected product number and original data information from the Land Observation Satellite Service Center.

8. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 1, characterized in that, The method further includes: The segmented map dataset is cropped and stitched together to obtain time-series SAR image data on the standard map sheet.

9. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 8, characterized in that, The method also includes importing the obtained time-series SAR image map dataset and standard map data into the Land Observation Satellite Service Center website for verification.

10. The automatic zoning method for domestic land exploration data based on spatial geometric overlap as described in claim 2, characterized in that, The threshold is 90%.

Images