Remote sensing image on-line analyzing and processing system

Abstract

The invention relates to the technical field of remote sensing, in particular to a remote sensing image on-line analysis processing system, which comprises an identification module used for acquiring a target remote sensing image and corresponding auxiliary geographic data; the correction module is used for correcting the target remote sensing image; the processing module is used for optimizing the correction result of the target remote sensing image; through multi-source data fusion, efficient storage, an adaptive correction model and a real-time error feedback mechanism, the problems of low precision, poor efficiency and insufficient complex scene adaptability in traditional remote sensing image processing are solved, high-precision and high-efficiency online analysis processing is realized, and the application value of the remote sensing image is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of remote sensing technology, and more specifically, to an online remote sensing image analysis and processing system. Background Technology

[0002] With the rapid development of remote sensing technology, remote sensing imagery has played a crucial role in geographic information acquisition, resource monitoring, and environmental assessment. However, geometric distortion and radiometric errors in remote sensing imagery remain key issues affecting the accuracy of its applications. Traditional remote sensing image processing systems typically rely on a single data source (such as optical imagery) for correction, making it difficult to effectively eliminate errors caused by complex terrain, building obstruction, and other factors. Furthermore, existing systems suffer from low computational efficiency, high storage redundancy, and poor adaptability of correction models when processing large-scale image data. Summary of the Invention

[0003] In view of this, the present invention addresses the shortcomings of the prior art by proposing an online remote sensing image analysis and processing system, aiming to solve at least one of the problems mentioned in the background art.

[0004] This invention provides an online remote sensing image analysis and processing system, comprising: an identification module for acquiring target remote sensing images and corresponding auxiliary geographic data; A correction module is used to correct the target remote sensing image; The processing module is used to optimize the correction results of the target remote sensing image.

[0005] In some embodiments, the identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, including: The target remote sensing image and auxiliary geographic data are stored in blocks and redundantly backed up.

[0006] In some embodiments, the auxiliary geographic data includes: vector map data, DEM data, and land use classification data. The vector map data is used to provide linear features as control points, the DEM data is used for terrain relief compensation, and the land use classification data is used to verify feature boundaries.

[0007] In some embodiments, the identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: performing radiometric correction and geometric coarse correction on the target remote sensing images, and aligning the DEM data and the vector map data with the coordinate system of the target remote sensing images.

[0008] In some embodiments, the identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: using the DEM data, obtaining the displacement caused by the elevation of each image point in the target remote sensing image through a collinearity equation, in order to eliminate image point position deviations caused by terrain undulations.

[0009] In some embodiments, the correction module is used to correct the target remote sensing image, including: The coordinates of prominent features are extracted from the vector map data as additional GCPs. The GCPs and the DEM data are combined, and the least squares method is used to obtain the correction transformation matrix to construct a geometric correction model.

[0010] In some embodiments, the correction module is used to correct the target remote sensing image, and further includes: The boundaries of the target features are extracted from the radar image corresponding to the target remote sensing image, and the target remote sensing image is further corrected.

[0011] In some embodiments, the processing module is used to optimize the correction result of the target remote sensing image, including: applying a geometric correction model to resample the optical image corresponding to the target remote sensing image to generate a preliminary corrected image, overlaying the preliminary corrected image with the vector map data, checking for deviations, and obtaining a first deviation result.

[0012] In some embodiments, the processing module is used to optimize the correction result of the target remote sensing image, and further includes: comparing the pre-corrected image with the land use classification data to obtain a second deviation result.

[0013] In some embodiments, the parameters of the geometric correction model are adjusted based on the first deviation result and the second deviation result.

[0014] Compared with existing technologies, the advantages of this invention are as follows: By fusing vector maps (providing control points), DEMs (compensating for terrain undulations), land use classification data (verifying feature boundaries), and radar imagery (extracting building boundaries), a more accurate geometric correction model is constructed, effectively eliminating image point displacement and distortion caused by factors such as terrain and building occlusion. Utilizing the geometric stability of radar imagery, target feature boundaries are extracted, further correcting local distortions in optical imagery and solving complex scene problems that are difficult to handle in traditional optical correction. Target remote sensing imagery and auxiliary geographic data are stored in blocks with redundant backups, improving data transmission efficiency and enhancing system fault tolerance, avoiding data loss or processing interruption due to single-point failures. Obvious features (such as roads and rivers) are extracted from the vector map as additional control points (GCPs), and combined with DEM data, the correction transformation matrix is ​​optimized using the least squares method, making the model more adaptable to complex terrain and feature variations. Image point displacement caused by elevation is calculated using DEM data, and projection differences caused by terrain undulations are eliminated through collinearity equations, significantly improving the correction accuracy of images in mountainous or hilly areas. After initial correction, the imagery is overlaid with the vector map to check for geometric deviations, and compared with land use classification data to verify the consistency of land feature boundaries, generating the first and second deviation results.

[0015] The above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure.

[0016] Other features and aspects of this disclosure will become clearer from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a functional block diagram of the online remote sensing image analysis and processing system provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0020] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0021] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0022] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0023] See Figure 1 As shown, an online remote sensing image analysis and processing system according to an embodiment of this application includes: The identification module is used to acquire remote sensing images of the target and the corresponding auxiliary geographic data; A correction module is used to correct the target remote sensing image; The processing module is used to optimize the correction results of the target remote sensing image.

[0024] In some specific embodiments, the identification module is used to acquire remote sensing images of the target and corresponding auxiliary geographic data, including: The target remote sensing image and auxiliary geographic data are stored in blocks and redundantly backed up.

[0025] It should be understood that the target remote sensing image and auxiliary geographic data (such as vector maps, DEMs, land use classification data) are divided into multiple regular blocks (such as grids or strips) according to spatial extent or data volume, and each block is stored independently. Based on the image resolution and coverage, the data is divided into blocks of fixed size (e.g., 512×512 pixels or 1°×1° area), and each block is assigned a unique identifier (such as row and column numbers or geographic coordinates).

[0026] For each data block, generate multiple copies (e.g., 2-3 copies) and store them in different physical locations (e.g., distributed servers, cloud storage nodes) or different storage media. Employ RAID technology, distributed file systems (e.g., HDFS), or cloud storage redundancy strategies (e.g., AWS S3 multi-availability zone backup) to ensure that data can still be recovered from other copies in the event of any single point of failure.

[0027] Users or systems only need to request chunked data of the target area, rather than loading the entire image, reducing network bandwidth consumption and transmission latency. After being segmented, multiple data chunks can be called simultaneously for parallel computation (such as geometric correction and resampling), improving processing speed. A redundant backup mechanism prevents data loss due to hardware failure, network interruption, or human error, ensuring continuous system operation. If a chunk of data is corrupted or lost, it can be directly restored from the backup copy without needing to reacquire the original data.

[0028] In some specific embodiments, the auxiliary geographic data includes: vector map data, DEM data, and land use classification data. The vector map data is used to provide linear features as control points, the DEM data is used for terrain relief compensation, and the land use classification data is used to verify feature boundaries.

[0029] It should be understood that extracting precise coordinates of linear features (such as roads, rivers, and building boundaries) from vector maps serves as additional control points (GCPs), supplementing the limited number of manual control points in traditional remote sensing imagery. Combined with DEM data, the geometric correction model is optimized using the least squares method to address image point displacement caused by complex terrain or feature variations.

[0030] Using the elevation information provided by the DEM, the displacement (projection difference) of each image point caused by terrain undulations is calculated through the collinearity equation, and this displacement is compensated in geometric correction. This is used to eliminate geometric distortions caused by terrain undulations such as mountains and hills, ensuring accurate alignment between optical imagery and vector maps.

[0031] The preliminarily corrected remote sensing images are overlaid and compared with land use classification data (such as vector boundaries of cultivated land, forest land, and building areas) to analyze the consistency of land feature boundaries.

[0032] By employing linear ground features (GCP) and terrain compensation, a more accurate geometric correction model is constructed, significantly reducing image point location errors across different terrains such as plains and mountains. Building boundaries extracted from radar imagery are further corrected to address local distortions in optical imagery caused by occlusion or projection differences. By comparing the corrected imagery with classification data, ground feature boundary deviations (such as confusion between farmland and construction land) are automatically detected, avoiding the subjectivity of manual visual inspection and improving the objectivity of the results.

[0033] In some specific embodiments, the identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: performing radiometric correction and geometric coarse correction on the target remote sensing images, and aligning the DEM data and the vector map data with the coordinate system of the target remote sensing images.

[0034] It should be understood that by using sensor parameters (such as gain and bias) or solar radiation models, the grayscale value deviations in the target remote sensing image caused by atmospheric scattering, sensor noise, etc., are corrected, and the pixel values ​​are converted into actual radiance or reflectance. Radiometric calibration formulas (such as Landsat's radiometric correction algorithm) or atmospheric correction models (such as FLAASH and 6S models) are used to eliminate the influence of the atmosphere and the sensor.

[0035] Geometric coarse correction, based on satellite orbital parameters and sensor attitude data (such as RPC parameters or ephemeris files), initially eliminates geometric distortions (such as tilt and rotation) in the image caused by factors such as satellite attitude and Earth curvature. Then, using the rational polynomial coefficients (RPC) provided by the satellite or a general sensor model (such as a rigorous physical model), the image coordinates are transformed to a map coordinate system (such as WGS84).

[0036] Force the coordinate systems of DEM data and vector map data (such as UTM, GCS) to be unified with the coordinate system of the target remote sensing image to ensure spatial consistency of multi-source data. Use coordinate transformation matrices (such as affine transformation, projection transformation) or resampling (such as bilinear interpolation) to register auxiliary data and images to the same geographic grid.

[0037] Radiometric correction eliminates atmospheric and sensor errors, ensuring that image grayscale values ​​reflect the true reflectance characteristics of ground features, providing a reliable foundation for subsequent classification and quantitative analysis (such as vegetation index calculation). Coarse geometric correction initially eliminates systematic geometric distortions (such as displacement caused by satellite side-viewing), reducing the computational load of subsequent fine correction and improving processing efficiency. Coordinate system alignment ensures spatial consistency between the DEM, vector map, and remote sensing image, avoiding data misalignment due to coordinate system differences (such as elevation mismatch with image points during terrain compensation). Block storage support allows for the storage of radiometrically and geometrically corrected images and auxiliary data in blocks, facilitating subsequent on-demand loading and parallel processing (such as terrain compensation and boundary verification).

[0038] In some specific embodiments, the identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: using the DEM data, obtaining the displacement caused by the elevation of each image point in the target remote sensing image through a collinearity equation, in order to eliminate image point position deviations caused by terrain undulations.

[0039] It should be understood that, based on the sensor imaging geometric model (such as central projection), the mathematical relationship between image point coordinates (x, y), ground point coordinates (X, Y, Z), and sensor parameters (such as focal length and attitude angle) is established through collinearity equations. For each image point in the image, combined with the elevation value Z in the DEM data, the three-dimensional coordinates (X, Y, Z) of its corresponding ground point are calculated, and then substituted into the collinearity equations to solve for the displacements Δx and Δy of the image point due to elevation differences.

[0040] The calculated displacements Δx and Δy are back-compensated into the original image coordinates to correct the image point position deviation caused by terrain undulations, so that the image meets the requirements of orthophoto projection.

[0041] By adjusting the pixel positions of the image through geometric transformations (such as affine transformations or polynomial corrections), a corrected image that eliminates the influence of terrain is generated.

[0042] By accurately calculating the elevation displacement of each pixel using DEM data, geometric distortion problems (such as "forward tilt" or "backward tilt") in complex terrain areas like mountains and hills are resolved, significantly improving orthorectification accuracy. Displacement compensation values ​​are dynamically adjusted for different elevation areas, avoiding error accumulation in areas with drastic terrain changes, as is common in traditional global correction models. After eliminating terrain displacement, the coordinate systems of remote sensing imagery are perfectly aligned with auxiliary data such as vector maps and land use classification data, ensuring the accuracy of subsequent analyses (such as feature boundary verification and change detection).

[0043] In some specific embodiments, the correction module is used to correct the target remote sensing image, including: The coordinates of prominent features are extracted from the vector map data as additional GCPs. The GCPs and the DEM data are combined, and the least squares method is used to obtain the correction transformation matrix to construct a geometric correction model.

[0044] It should be understood that precise geographic coordinates (such as latitude, longitude, or projected coordinates) of obvious linear features or characteristic points are extracted from vector map data (such as roads, rivers, building boundaries, etc.) and used as additional ground control points (GCPs) to supplement traditionally manually selected GCPs. Spatial analysis algorithms (such as buffer analysis and topology matching) are used to automatically identify features in the vector map that match the characteristics of remote sensing imagery (such as road intersections and bridges) and extract their coordinates as GCPs.

[0045] Traditional GCPs are used to establish a preliminary mapping relationship between image coordinates and geographic coordinates. The displacement (Δx, Δy) of each GCP due to elevation differences is calculated using collinearity equations, and its coordinates are corrected to reflect the influence of terrain undulations. The original coordinates of the GCPs and the DEM-compensated coordinates are substituted into a geometric correction model (such as a polynomial transformation or rational function model), and the optimal transformation matrix (such as an affine transformation matrix or RPC parameters) is fitted using the least squares method to minimize the sum of squared errors between the image points and geographic coordinates.

[0046] Based on the optimized transformation matrix, a mathematical mapping relationship is established from image coordinates (x, y) to geographic coordinates (X, Y), generating a geometrically corrected orthophoto. The original image is resampled in grayscale (e.g., by bilinear interpolation or cubic convolution), and the pixel positions are adjusted according to the transformation matrix. Terrain compensation is then performed on each pixel using DEM data to ensure the correction results meet orthophoto projection requirements.

[0047] In some specific embodiments, the correction module is used to correct the target remote sensing image, and further includes: The boundaries of the target features are extracted from the radar image corresponding to the target remote sensing image, and the target remote sensing image is further corrected.

[0048] It should be understood that radar imagery (such as InSAR and optical-radar fusion data) is sensitive to targets such as buildings and terrain. Edge detection algorithms (such as the Canny operator and gradient segmentation) or classification models (such as object-oriented segmentation and deep learning) are used to extract the boundaries of these targets (e.g., building outlines and steep terrain lines). Based on the geometric characteristics of radar imagery (e.g., building displacement caused by side-view imaging) and grayscale features (e.g., the difference in reflection between buildings and surrounding areas), the boundaries of these targets are identified. Combining DEM data to correct geometric distortions in radar imagery (e.g., scaling effects caused by terrain undulations) improves the accuracy of boundary extraction.

[0049] Registration and correction of radar boundaries with optical imagery: The ground feature boundaries extracted from the radar imagery are used as additional constraints and jointly optimized with the geometric correction model of the optical imagery (based on GCP and DEM). The radar boundaries are projected onto the optical imagery coordinate system and compared with the corresponding ground feature boundaries in the optical imagery to calculate deviations (such as building corner offsets). Based on the deviations, the parameters of the geometric correction model (such as local affine transformations or polynomial coefficients) are adjusted to further correct the image point positions in the optical imagery.

[0050] Dynamic compensation and iterative optimization: To address the local mismatch between radar and optical imagery caused by differences in observation angle and time, the correction model is dynamically adjusted using iterative weighted least squares or elastic registration algorithms, prioritizing the correction of errors in boundary regions. Boundary regions (such as densely built-up areas) are segmented and assigned higher weights to ensure their geometric accuracy is superior to that of smooth regions. The reliability of the correction results is verified through cross-validation of multi-temporal data using redundant backup radar imagery.

[0051] In some specific embodiments, the processing module is used to optimize the correction result of the target remote sensing image, including: applying a geometric correction model to resample the optical image corresponding to the target remote sensing image to generate a preliminary corrected image, overlaying the preliminary corrected image with the vector map data, checking for deviations, and obtaining a first deviation result.

[0052] It should be understood that the geometric correction model (such as an affine transformation matrix or a rational function model) built based on the correction module resamples the optical image of the target remote sensing image, mapping each pixel to the correct geographic coordinates. The resampling method is bilinear interpolation, which calculates the new pixel value by weighting the gray values ​​of the surrounding four pixels, balancing accuracy and efficiency.

[0053] The preliminary corrected imagery is spatially overlaid with vector map data (such as roads, building boundaries, and waterways). Geometric deviations are detected by comparing the image pixel locations with the geographic coordinates of the vector features. The coordinates of prominent features in the imagery (such as road intersections and bridges) and their corresponding points in the vector map are extracted, and the offset is calculated. The boundaries of features in the imagery (such as building outlines) are overlaid with the boundaries of the vector map, and the mean and variance of the edge deviations are calculated.

[0054] Based on the results of the overlay analysis, the first deviation result is generated, including global error statistics (such as average offset and maximum offset) and local error distribution maps (such as heat maps or vector error vectors).

[0055] In some specific embodiments, the processing module is used to optimize the correction result of the target remote sensing image, and further includes: comparing the pre-corrected image with the land use classification data to obtain a second deviation result.

[0056] In some specific embodiments, the parameters of the geometric correction model are adjusted based on the first deviation result and the second deviation result.

[0057] It should be understood that spatial overlay of preliminary corrected images with land use classification data (such as vectorized land type boundaries or raster classification results) is used to detect geometric deviations by comparing the geographic coordinates of image pixel locations with those of classification boundaries.

[0058] Deviation detection method: Extract the boundaries of land features (such as farmland, forest, and building areas) from the image and overlay them with the boundaries in the land use classification data to calculate the offset.

[0059] The geometric correction model parameters are adjusted by combining the first and second bias results: the first bias result reflects the geometric deviation between the imagery and the vector map (e.g., roads, buildings), focusing on the local accuracy of prominent features. The second bias result reflects the global consistency between the imagery and land use classification data, focusing on the overall registration of feature boundaries. Weights are assigned to the two types of bias results according to the application scenario requirements (e.g., prioritizing the first bias for urban areas and the second bias for agricultural monitoring). Global model parameters (e.g., polynomial transform coefficients) are fine-tuned to reduce overall offset.

[0060] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A remote sensing image online analysis and processing system, characterized in that, include: The identification module is used to acquire remote sensing images of the target and corresponding auxiliary geographic data; A correction module is used to correct the target remote sensing image; The processing module is used to optimize the correction results of the target remote sensing image.

2. The remote sensing image online analysis and processing system according to claim 1, characterized in that, The identification module is used to acquire remote sensing images of the target and corresponding auxiliary geographic data, including: The target remote sensing image and auxiliary geographic data are stored in blocks and redundantly backed up.

3. The remote sensing image online analysis and processing system according to claim 2, characterized in that, The auxiliary geographic data includes: vector map data, DEM data, and land use classification data. The vector map data is used to provide linear features as control points, the DEM data is used for terrain relief compensation, and the land use classification data is used to verify feature boundaries.

4. The remote sensing image online analysis and processing system according to claim 3, characterized in that, The identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: performing radiometric correction and geometric coarse correction on the target remote sensing images, and aligning the DEM data and the vector map data with the coordinate system of the target remote sensing images.

5. The remote sensing image online analysis and processing system according to claim 4, characterized in that, The identification module is used to acquire target remote sensing images and corresponding auxiliary geographic data, and further includes: using the DEM data, obtaining the displacement caused by the elevation of each image point in the target remote sensing image through the collinearity equation, in order to eliminate the image point position deviation caused by terrain undulation.

6. The online remote sensing image analysis and processing system according to claim 5, characterized in that, The correction module is used to correct the target remote sensing image, including: The coordinates of prominent features are extracted from the vector map data as additional GCPs. The GCPs and the DEM data are combined, and the least squares method is used to obtain the correction transformation matrix to construct a geometric correction model.

7. The online remote sensing image analysis and processing system according to claim 6, characterized in that, The correction module is used to correct the target remote sensing image, and further includes: The boundaries of the target features are extracted from the radar image corresponding to the target remote sensing image, and the target remote sensing image is further corrected.

8. The remote sensing image online analysis and processing system according to claim 7, characterized in that, The processing module is used to optimize the correction result of the target remote sensing image, including: applying a geometric correction model to resample the optical image corresponding to the target remote sensing image to generate a preliminary corrected image, overlaying the preliminary corrected image with the vector map data, checking the deviation, and obtaining a first deviation result.

9. The remote sensing image online analysis and processing system according to claim 8, characterized in that, The processing module is used to optimize the correction results of the target remote sensing image, and also includes: comparing the pre-corrected image with the land use classification data to obtain a second deviation result.

10. The online remote sensing image analysis and processing system according to claim 9, characterized in that, The parameters of the geometric correction model are adjusted based on the first deviation result and the second deviation result.

Images