An infrared remote sensing control point image database construction method and system

Abstract

The application provides an infrared remote sensing control point image database construction method and system, and relates to the technical field of remote sensing image processing. The method comprises the following steps: acquiring a reference image; extracting image feature points of the reference image to form an initial control point set; performing coordinate conversion on the initial control points in the initial control point set to determine a geographic control point set; screening each geographic control point in the geographic control point set by using a random sample consensus algorithm to determine a target control point set; naming and encoding each target control point in the target control point set according to a predetermined rule to obtain an encoded control point set; taking each encoded control point in the encoded control point set as the center, performing multi-scale window clipping in a to-be-matched infrared image and a reference visible light image respectively to obtain a control point image block; and combining the control point image block and the attribute information corresponding to the encoded control point set to construct an infrared remote sensing control point image database.

Description

Technical Field

[0001] This invention relates to the field of remote sensing image processing technology, and in particular to a method and system for constructing an infrared remote sensing control point image database. Background Technology

[0002] With the continuous advancement of remote sensing technology and the significant improvement in sensor performance, remote sensing image data is being widely applied in several key fields, including geological and mineral exploration, dynamic environmental monitoring, disaster early warning and forecasting, detailed urban mapping, and military reconnaissance. To extract accurate geospatial information from these images, geometrical correction (i.e., orthorectification) is usually necessary.

[0003] Infrared remote sensing imagery differs fundamentally from visible light imagery due to its unique imaging mechanism. It is more sensitive to differences in temperature radiation of ground features, often exhibiting low contrast and uneven grayscale dynamic range distribution. Currently available global control point databases and high-precision control point extraction methods are primarily designed for visible light or multispectral imagery, lacking specific reference data support for infrared remote sensing imagery. In critical applications such as emergency remote sensing, while infrared satellites can penetrate smoke and effectively monitor fires, the lack of high-precision infrared control point reference data severely limits their ability to perform rapid geometric correction and accurate positioning.

[0004] Traditional geometric correction methods primarily rely on data such as ground control points and digital elevation models (DEMs). This requires acquiring sufficient control points and corresponding elevation information before importing them into a specialized remote sensing processing system for correction. However, images acquired by different sensors often differ in resolution, coordinate system, and imaging time-phase characteristics. This places diverse demands on control point reference data, posing significant challenges to the storage and management of control point information. Under current technological frameworks, control points are often stored piecemeal using a combination of measured field coordinates and image labels, lacking systematic integration and unified management. This results in low data utilization efficiency and often necessitates periodic re-collection. Therefore, constructing a database capable of efficiently organizing, managing, and conveniently accessing control point information is particularly urgent. Summary of the Invention

[0005] To address the problem that, under the existing technological framework, control points are often stored in a scattered manner using a combination of field-measured coordinates and image labels, lacking systematic integration and unified management, resulting in low data utilization efficiency and the need for periodic re-collection, this invention provides a method and system for constructing an infrared remote sensing control point image database.

[0006] The technical solutions provided by the embodiments of the present invention are as follows: The first aspect of this invention provides a method for constructing an infrared remote sensing control point image database, comprising: S1: Acquire baseline image; S2: Extract image feature points from the reference image to form an initial set of control points; S3: Perform coordinate transformation on the initial control points in the initial control point set to determine the geographic control point set; S4: The set of geographic control points is determined by filtering each geographic control point in the set of geographic control points using a random sampling consensus algorithm; S5: According to predetermined rules, name and encode each target control point in the target control point set to obtain the coded control point set; S6: Using each coded control point in the coded control point set as the center, perform multi-scale window cropping in the infrared image to be matched and the reference visible light image respectively to obtain control point image blocks; S7: Combine the attribute information corresponding to the control point image blocks and the coded control point set to construct an infrared remote sensing control point image database.

[0007] A second aspect of the present invention provides an infrared remote sensing control point image database construction system, comprising: processor; The memory stores computer-readable instructions, which, when executed by a processor, implement the infrared remote sensing control point image database construction method as described in the first aspect.

[0008] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, the infrared remote sensing image control point image database constructed by this invention achieves unified and standardized management of infrared satellite image control point data, providing reliable support for rapid geometric correction of infrared images. Compared with the existing technology that lacks infrared control point reference data, this invention establishes for the first time a nationwide infrared control point image database, clarifying the registration relationship between infrared images and reference images, greatly improving the accuracy and efficiency of geometric correction of infrared remote sensing images. This database can meet the needs of daily remote sensing mapping and processing and emergency remote sensing for sudden events, and has broad application prospects and important practical value in applications such as precise positioning of remote sensing images, disaster monitoring, and emergency services. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1This is a flowchart illustrating a method for constructing an infrared remote sensing control point image database, as provided in an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of a method for constructing an infrared remote sensing control point image database according to an embodiment of the present invention.

[0012] Figure 3 This is a schematic diagram of a control point coding rule provided in an embodiment of the present invention.

[0013] Figure 4 This is a database system architecture diagram provided for an embodiment of the present invention.

[0014] Figure 5 This is a schematic diagram of the structure of an infrared remote sensing control point image database construction system provided in an embodiment of the present invention. Detailed Implementation

[0015] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0016] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0017] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0018] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] Reference manual attached Figure 1 The diagram illustrates a flowchart of an infrared remote sensing control point image database construction method provided by an embodiment of the present invention.

[0021] Reference manual attached Figure 2The diagram shows a structural schematic of an infrared remote sensing control point image database construction method provided by an embodiment of the present invention.

[0022] This invention provides a method for constructing an infrared remote sensing control point image database. This method can be implemented using an infrared remote sensing control point image database construction device, which can be a terminal or a server. The processing flow of the infrared remote sensing control point image database construction method may include the following steps: S1: Acquire the baseline image.

[0023] Among them, the reference image refers to the set of data images that serve as the basis for subsequent control point extraction, coordinate transformation, and multi-source image matching, providing a unified spatial reference and feature extraction basis.

[0024] It should be noted that by acquiring infrared multispectral images and high-resolution visible light remote sensing images as reference images, the thermal radiation characteristics of infrared images and the texture detail features of visible light images can be utilized simultaneously to improve the accuracy and stability of subsequent control point extraction and cross-modal matching, providing a reliable data foundation for the construction of an infrared remote sensing control point image database.

[0025] In one possible implementation, the reference imagery specifically includes: infrared multispectral imagery and high-resolution visible light remote sensing imagery.

[0026] Specifically, the reference images include infrared remote sensing images from the HJ-2 A / B satellites with a resolution of 96m and visible light remote sensing images from Google Earth with a resolution of 19.05m. The remote sensing images achieve full coverage of the target area across the country through multi-scene image stitching.

[0027] In one possible implementation, the process after S1 and before S2 includes: S1A: Preprocess the infrared multispectral image in the reference image.

[0028] S1A specifically refers to: By combining the probability density function of infrared multispectral images, grayscale distribution mapping processing is performed on the infrared multispectral images:

[0029] Where r represents the grayscale value in the infrared multispectral image. Let represent the probability density function of the infrared multispectral image, w represent the intermediate gray level variable in the infrared image, and d represent the differential sign. Let z represent the probability density function of the visible light image of the target, z represent the gray value in the target image, and t represent the intermediate gray value variable in the target image.

[0030] It should be noted that before matching, the infrared image needs to be histogram-defined to make its grayscale distribution as close as possible to that of the visible light image, which can effectively reduce the matching failure rate caused by spectral differences.

[0031] S2: Extract image feature points from the reference image to form an initial set of control points.

[0032] Image feature points refer to key pixels in an image that exhibit significant grayscale variations, texture changes, or edge structures. They are typically represented as corner points, intersections, or areas with prominent local textures, and are used to characterize stable spatial features in the image. The initial control point set refers to a set of candidate control points with corresponding relationships obtained after performing feature point extraction and cross-image matching on the reference image. These are used for subsequent coordinate transformation and error filtering.

[0033] It should be noted that by extracting image feature points from the reference image and forming an initial set of control points, stable and identifiable key areas in the image can be effectively obtained, improving the matching accuracy between subsequent infrared and visible light images and the positioning accuracy of control points, thus providing a reliable foundation for subsequent geographic coordinate transformation and control point selection.

[0034] In one possible implementation, S2 specifically includes: S201: Calculate the autocorrelation matrix of the reference image pixels:

[0035] Where M represents the autocorrelation matrix, I represents the Gaussian smoothing window function. x I represents the gradient of the image in the x-direction. y This represents the gradient of the image in the y-direction.

[0036] S202: Calculate the eigenvalues ​​of the autocorrelation matrix and define the corner response function based on the eigenvalues:

[0037] Where R represents the corner response function, det() represents the determinant of the matrix, k represents the empirical coefficient of the corner response, and trace() represents the trace of the matrix. and All of these represent the eigenvalues ​​of the autocorrelation matrix.

[0038] S203: When the corner response function value is greater than the threshold, determine the candidate feature points.

[0039] S204: Secondary filtering of candidate feature points using the Forstner operator:

[0040] Where w represents the isotropy of the Forstner operator and q represents the roundness of the Forstner operator.

[0041] S205: Based on the results of the second screening, candidates for high robustness control points are determined.

[0042] S206: By using an improved scale-invariant feature transformation algorithm, feature descriptions are constructed for candidates with high robustness of control points, resulting in cross-scale feature descriptors.

[0043] S207: Based on cross-scale feature descriptors and mutual information constraints, similarity matching is performed on control point candidates in the reference image and the reference image to determine the initial set of control points.

[0044]

[0045]

[0046]

[0047] in, Indicating infrared feature points With visible light feature points The overall matching cost between them This represents the SIFT feature difference weighting coefficient. Indicates the infrared SIFT descriptor. This represents the visible light SIFT descriptor. Represents the mutual information weighting coefficient. Indicates infrared window With visible light window Mutual information between them This represents the information entropy of the infrared window. This represents the information entropy of the visible light window. This represents the joint entropy of the infrared window and the visible light window. This represents the probability of grayscale value x in the infrared window. This represents the joint probability of infrared grayscale x and visible light grayscale y.

[0048] Specifically, since infrared and visible light images are heterogeneous images, traditional grayscale correlation matching is prone to failure. This embodiment adopts a combined strategy of "Forstner / Harris corner detection + MI-SIFT descriptor". Furthermore, in terms of matching strategy, the Fitting Global Transform model is used, with a maximum matching tolerance of 2 pixels. One HJ-2 image is overlapped and matched with 4-5 Google Earth images. Through batch processing, approximately 60,000 initial control points are extracted across the entire region.

[0049] It should be noted that this step, by combining Harris corner detection, Forstner operator filtering, improved SIFT feature description, and mutual information constraint matching, can extract control point features from infrared and visible light images that are highly stable, noise-resistant, and adaptable across scales. This not only improves the feature matching accuracy and robustness between multi-source remote sensing images, but also enhances the consistency and reliability of control point identification in complex scenarios, providing a high-quality initial control point foundation for subsequent geographic coordinate transformation and database construction.

[0050] S3: Perform coordinate transformation on the initial control points in the initial control point set to determine the geographic control point set.

[0051] The geographic control point set refers to a set of control point data that includes image pixel locations, geographic coordinates, and image correspondences after being mapped to a unified geographic coordinate system. It is used for subsequent control point selection and spatial matching processing.

[0052] Specifically, the control point coordinate transformation step uses Python to call the GDAL library interface to convert the row and column numbers of the acquired control points on the image into latitude and longitude coordinates in the WG84 geographic coordinate system, thereby obtaining two sets of coordinate data: control point coordinates in infrared imagery and control point coordinates in visible light imagery.

[0053] In one possible implementation, S3 specifically includes: S301: Obtain the pixel coordinates of infrared image control points and visible light image control points in the initial control point set.

[0054] S302: Read the GeoTransform parameters corresponding to the infrared multispectral image and the high-resolution visible light remote sensing image through the GDAL library to determine the affine transformation matrix parameters of the image.

[0055] S303: Based on the affine transformation matrix parameters, perform coordinate transformation on the pixel coordinates of control points in the infrared image and the visible light image to obtain the geographic coordinates.

[0056] Where X represents the geographic horizontal coordinate, Y represents the geographic vertical coordinate, P represents the pixel column number, L represents the pixel row number, a represents the geographic horizontal coordinate corresponding to the top-left pixel, b represents the geographic resolution corresponding to the pixel width, c represents the row rotation parameter, d represents the geographic vertical coordinate corresponding to the top-left pixel, e represents the column rotation parameter, and f represents the geographic resolution corresponding to the pixel height.

[0057] S304: Map the pixel coordinates of infrared image control points and visible light image control points to a unified geographic coordinate system based on geographic coordinates.

[0058] in, This represents the i-th ground control point. Represents the pixel coordinates in the infrared image. Represents the column coordinates in the infrared image. Represents the row coordinates in the infrared image. Represents latitude and longitude coordinates in the WGS84 coordinate system. Indicates latitude, Indicates longitude.

[0059] S305: Based on the mapping results, generate a set of geographic control points containing the row and column numbers of control points, geographic coordinates, and image correspondences.

[0060] It should be noted that this step, by reading the GeoTransform parameters of the remote sensing image and performing a unified geographic coordinate system mapping, can uniformly transform the control points in the infrared and visible light images to the same spatial reference frame. This effectively improves the spatial positioning consistency and coordinate matching accuracy among multi-source remote sensing images, while also facilitating subsequent control point selection, spatial distribution analysis, and standardized database management, thereby enhancing the reusability of control point data and the reliability of engineering applications.

[0061] S4: The target control point set is determined by filtering the geographic control points in the set of geographic control points through the random sampling consensus algorithm.

[0062] The random sampling consensus algorithm is a robust model estimation algorithm used to remove erroneous matching points. It constructs a transformation model by randomly selecting some matching points and filters out interior points that meet the consistency conditions based on projection errors. The target control point set refers to a set of highly reliable control points obtained after error filtering, spatial homogenization, and outlier removal, which is used for subsequent control point coding and image block cropping.

[0063] Specifically, the RANSAC algorithm is first applied to remove mismatched control point pairs. Then, the quality is evaluated based on the uniformity of the distribution of control points within the region. Approximately 50% of the remaining control points are selected and retained according to the uniform distribution quality screening method to ensure that the final control points are evenly distributed in space, resulting in approximately 30,000 control points.

[0064] In one possible implementation, S4 specifically includes: S401: Based on the set of geographic control points, determine the matching point pairs between the infrared image control points and the reference image control points, and establish the initial affine transformation model between the matching point pairs:

[0065] Where (x,y) represents the coordinates of a point in the original image, , () represents the transformed target coordinates. Indicates the rotation parameter in the x-direction. Indicates the shearing parameter in the x-direction. This represents the translation parameter in the x-direction. This represents the shear parameter in the y-direction. This represents the rotation parameter in the y-direction. This represents the translation parameter in the y-direction.

[0066] S402: Based on the initial affine transformation model, randomly select a preset number of matching point pairs from the matching point pairs and calculate the temporary affine transformation model.

[0067] S403: Based on the temporary affine transformation model, perform a projection transformation on the remaining matching point pairs and calculate the projection error of each matching point pair: .

[0068] S404: Determine if the projection error is less than the error threshold. If yes, identify the corresponding matching point pair as an interior point and add it to the current consistency control point set, then proceed to step S405. Otherwise, identify the corresponding matching point pair as an exterior point and perform a removal mark.

[0069] S405: Divide the target area into multiple rule grids, and assign each control point to the corresponding rule grid according to the geographical coordinates of each control point in the current set of consistent control points.

[0070] S406: Calculate the spatial distribution characteristics of control points within each regular grid.

[0071] S407: When the number of control points within a regular grid exceeds a preset threshold, calculate the comprehensive evaluation value based on the spatial distribution characteristics of the control points.

[0072] Where p represents the control point to be evaluated. This represents the matching similarity weight coefficient. This represents the similarity score of the matching of control point p. This represents the current control point p to be evaluated and its neighboring control points. The distance between them.

[0073] S408: Based on the comprehensive evaluation value, sort the control points within the regular grid, and determine the set of spatially homogenized control points based on the sorting results.

[0074] S409: Determine the set of spatially homogenized control points as the target control point set.

[0075] It should be noted that by constructing an affine transformation model and combining it with a random sampling consensus algorithm to screen control points for errors, erroneous matching points and outliers can be effectively eliminated, improving the accuracy and stability of control point matching results. At the same time, through regular grid division and spatial homogenization evaluation mechanism, the problem of local clustering of control points can be avoided, enhancing the spatial distribution balance of control points, thereby improving the overall reliability and robustness of subsequent remote sensing image registration, spatial positioning, and database construction processes.

[0076] Reference manual attached Figure 3 The diagram illustrates a control point coding rule provided by an embodiment of the present invention.

[0077] S5: According to predetermined rules, name and encode each target control point in the target control point set to obtain the coded control point set.

[0078] Naming and coding refers to the process of assigning a unique identifier to each target control point according to a preset coding rule. This is used to distinguish different control points and establish a unified data management method. The coded control point set refers to the control point data set after naming and coding, which includes the unique code, spatial location, and associated attribute information of each control point. This data is used for subsequent image block organization and database storage management.

[0079] Specifically, the naming rule for control points is “SRRCCrrccYYNNNN”, where S represents the control point type code (0 / 1 for visible / infrared control points), RR and CC are the row and column numbers of the large grid to which the control point belongs (390km×390km), rr and cc are the row and column numbers of the sub-grids divided by the large grid (10km×10km, the large grid is divided into 39×39 small grids), YY are the last two digits of the extraction year, and NNNN is the sequential number (locally increasing) of the control points within the large grid.

[0080] In one possible implementation, S5 specifically includes: S501: Based on the geographical coordinates of each target control point in the target control point set, the target area is divided into regular grids to determine the large grid area and small grid area to which each target control point belongs.

[0081] S502: Determine the corresponding control point type code, grid number, and year identifier based on the image data source, grid region, and image acquisition time for each target control point.

[0082] S503: Generate the corresponding sequence number based on the generation order of each target control point within its respective small grid area.

[0083] S504: Combine the control point type code, grid number, year identifier, and serial number to generate the control point code.

[0084] S505: Establish a spatial index relationship between control point codes and the geographical locations of target control points to obtain a set of coded control points.

[0085] Specifically, the encoding format is: SRRCCrrccYYNNNN (15 digits in total). S (1 digit): Data source / type code. Infrared image control points are set to "1". RRCC (4 digits): Large grid coordinates. Based on national standard grading or a custom 390km×390km grid. rrcc (4 digits): Small grid coordinates. The large grid is further subdivided into 39×39 small grids (10km×10km). This allows the index to directly locate local areas with high precision, greatly improving database retrieval speed. YY (2 digits): Year identifier. For example, "25" represents 2025, facilitating version control and time-phase updates. NNNN (4 digits): Serial number. A unique ID generated according to the acquisition order within the same small grid. This encoding is not only an ID but also a spatial index; the database can directly deduce the approximate location of the control point from the ID without reading the coordinate field.

[0086] It should be noted that by combining the grid region to which the control point belongs, the image source, and the acquisition time to generate a unified control point code, a unique identifier relationship with spatial location characteristics and temporal attribute characteristics can be established for each control point. This not only improves the standardization of control point data organization and management, but also enhances the efficiency and traceability in the process of control point retrieval, spatial index positioning, and database expansion and maintenance.

[0087] S6: Using each coded control point in the coded control point set as the center, perform multi-scale window cropping in the infrared image to be matched and the reference visible light image respectively to obtain control point image blocks.

[0088] Multi-scale window cropping refers to a processing method that uses control points as the center and cuts out image regions according to window ranges of different sizes to preserve local feature information at different scales around the control points. Control point image blocks refer to local image regions cropped from infrared and visible light images around coded control points, used for subsequent database storage, feature analysis, and image matching processing.

[0089] Specifically, in the control point image block cropping step, the ambient infrared image is cropped with the control point as the center to obtain an infrared control point image block of size 65×65 pixels, and the visible light reference image is cropped with the corresponding control point as the center to obtain a visible light image block of size 327×327 pixels. All feature information near the control point is retained for wide-span image feature matching.

[0090] In one possible implementation, S6 specifically includes: S601: Obtain the geographic coordinates and image pixel coordinates corresponding to each coded control point in the coded control point set.

[0091] S602: Determine the center position of each coding control point based on geographic coordinates and image pixel coordinates.

[0092] S603: Using the center position as the center, crop an infrared image window of a preset size from the infrared multispectral image to be matched, and crop a visible light image window of a preset size from the high-resolution visible light remote sensing image to obtain a dual-scale control point image block.

[0093] S604: Based on the resolution ratio between the infrared multispectral image to be matched and the reference visible light image, a consistency constraint is imposed on the physical coverage of the infrared image window and the visible light image window.

[0094] S605: Based on the consistency constraint results, determine whether the dual-scale control point image patch exceeds the corresponding image boundary. If so, fill the area exceeding the boundary with background values. Otherwise, retain the original image pixel values ​​and generate control point image patches of uniform size.

[0095] in, Indicates the cropped image block in Pixel value at the location, Indicates the location of the original image Pixel value at that location, This indicates that the current cropping position is located within the original image area. This indicates that the current cropping position is outside the original image area. This represents the background fill value.

[0096] Specifically, to support subsequent template matching, the database needs to store image chips centered on control points. This step performs dual-scale cropping: Infrared image patch: 65×65 pixels centered at GCP coordinates. This size covers an area of ​​approximately 6km×6km (at 96m resolution), sufficient to include significant terrain features.

[0097] Visible light image patch (reference benchmark): 327×327 pixels cropped from the same location in the Google Earth image.

[0098] Size design principle: 312 / 65≈5.0, which approximately corresponds to the resolution ratio of 19.05m to 96m. Maintaining consistency in physical coverage is crucial for multi-scale pyramid matching.

[0099] It should be noted that by performing dual-scale window cropping on infrared and visible light images centered on coded control points, and combining this with physical coverage consistency constraints based on resolution ratios, the spatial correspondence and scale consistency of local areas between different modal images can be effectively guaranteed. At the same time, through boundary detection and background value filling mechanisms, control point image blocks of uniform size can be generated, improving the standardization and stability of subsequent database storage, batch retrieval, and cross-modal image matching processing.

[0100] Reference manual attached Figure 4 The diagram illustrates a database system architecture provided by an embodiment of the present invention.

[0101] S7: Combine the attribute information corresponding to the control point image blocks and the coded control point set to construct an infrared remote sensing control point image database.

[0102] Among them, attribute information refers to descriptive data such as geographic coordinates, coding information, image source, time information, and spatial index relationships corresponding to control points. The infrared remote sensing control point image database is a structured data set used to store control point image blocks and their associated attribute information, for subsequent control point retrieval, image matching, and remote sensing analysis applications.

[0103] Specifically, the database stores control point attribute information including control point code, image data source, image resolution, cropping size, control point geographic coordinates, and matching error assessment, which are stored in association with control point image blocks in the form of CSV files or data tables.

[0104] In one possible implementation, the infrared remote sensing control point image database specifically includes: a physical storage layer, a spatial database management layer, a metadata organization layer, and a retrieval application layer.

[0105] The physical storage layer deploys the entire PostgreSQL data directory on a portable hard drive or other removable media. After installing the same version of PostgreSQL and PostGIS environment on the target computer, the database instance can be directly activated by mounting the same data directory, enabling cross-terminal database migration and reuse. It also features cloud capabilities.

[0106] It should be noted that by combining control point image blocks with corresponding attribute information to construct an infrared remote sensing control point image database, it is possible to achieve unified organization and associated management of control point data and image data, thereby improving the retrieval efficiency of control point information, data standardization, and the convenience of application in subsequent cross-modal remote sensing image matching and analysis.

[0107] In this embodiment of the invention, the infrared remote sensing image control point image database constructed by this invention achieves unified and standardized management of infrared satellite image control point data, providing reliable support for rapid geometric correction of infrared images. Compared with the existing technology that lacks infrared control point reference data, this invention establishes for the first time a nationwide infrared control point image database, clarifying the registration relationship between infrared images and reference images, greatly improving the accuracy and efficiency of geometric correction of infrared remote sensing images. This database can meet the needs of daily remote sensing mapping and processing and emergency remote sensing for sudden events, and has broad application prospects and important practical value in applications such as precise positioning of remote sensing images, disaster monitoring, and emergency services.

[0108] Reference manual attached Figure 5 The diagram shows a structural schematic of an infrared remote sensing control point image database construction system provided by the present invention.

[0109] The present invention also provides an infrared remote sensing control point image database construction system 20, applied to the above-mentioned infrared remote sensing control point image database construction method, comprising: Processor 201.

[0110] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, the infrared remote sensing control point image database construction method as described in the method embodiment is implemented.

[0111] The infrared remote sensing control point image database construction system 20 provided by the present invention can execute the above-described infrared remote sensing control point image database construction method and achieve the same or similar technical effects. To avoid duplication, the present invention will not elaborate further.

[0112] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0113] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0114] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0115] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0116] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0117] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0118] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0119] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0120] In the embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0121] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0123] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0124] This invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the infrared remote sensing control point image database construction method as described in the method embodiment.

[0125] The present invention provides a computer-readable storage medium that can implement the steps and effects of the infrared remote sensing control point image database construction method of the above method embodiments. To avoid repetition, the present invention will not repeat them.

[0126] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0127] The following points need to be explained: (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0128] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0129] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0130] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for constructing an infrared remote sensing control point image database, characterized in that, include: S1: Acquire baseline image; S2: Extract image feature points from the reference image to form an initial control point set; S3: Perform coordinate transformation on the initial control points in the initial control point set to determine the geographic control point set; S4: The set of geographic control points is determined by filtering each geographic control point in the set of geographic control points using a random sampling consensus algorithm; S5: According to predetermined rules, each target control point in the target control point set is named and encoded to obtain an encoded control point set; S6: Using each coded control point in the set of coded control points as the center, perform multi-scale window cropping in the infrared image to be matched and the reference visible light image respectively to obtain control point image blocks; S7: Combine the attribute information corresponding to the control point image blocks and the coded control point set to construct an infrared remote sensing control point image database.

2. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, The reference images specifically include: infrared multispectral images and high-resolution visible light remote sensing images.

3. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, The following is included after S1 and before S2: S1A: Preprocess the infrared multispectral image in the reference image; Specifically, S1A is: Based on the probability density function of the infrared multispectral image, grayscale distribution mapping processing is performed on the infrared multispectral image: ； Where r represents the grayscale value in the infrared multispectral image. Let represent the probability density function of the infrared multispectral image, w represent the intermediate gray level variable in the infrared image, and d represent the differential sign. Let z represent the probability density function of the visible light image of the target, z represent the gray value in the target image, and t represent the intermediate gray value variable in the target image.

4. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, S2 specifically includes: S201: Calculate the autocorrelation matrix of the reference image pixels; S202: Calculate the eigenvalues ​​of the autocorrelation matrix and define the corner response function based on the eigenvalues; S203: When the corner response function value is greater than the threshold, determine the candidate feature points; S204: The candidate feature points are further filtered using the Forstner operator; S205: Based on the results of the second screening, determine the candidates for high robustness control points; S206: By using an improved scale-invariant feature transformation algorithm, feature descriptions are constructed for the highly robust control point candidates to obtain cross-scale feature descriptors; S207: Based on the cross-scale feature descriptor and combined with mutual information constraints, perform similarity matching on the control point candidates in the reference image and the reference image to determine the initial control point set: ； ； ； ； in, Indicating infrared feature points With visible light feature points The overall matching cost between them This represents the SIFT feature difference weighting coefficient. Indicates the infrared SIFT descriptor, This represents the visible light SIFT descriptor. Represents the mutual information weighting coefficient. Indicates infrared window With visible light window Mutual information between them This represents the information entropy of the infrared window. This represents the information entropy of the visible light window. This represents the joint entropy of the infrared window and the visible light window. This represents the probability of grayscale value x in the infrared window. This represents the joint probability of infrared grayscale x and visible light grayscale y.

5. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, S3 specifically includes: S301: Obtain the pixel coordinates of infrared image control points and visible light image control points in the initial control point set; S302: Read the GeoTransform parameters corresponding to the infrared multispectral image and the high-resolution visible light remote sensing image through the GDAL library, and determine the image affine transformation matrix parameters; S303: Based on the image affine transformation matrix parameters, perform coordinate transformation on the pixel coordinates of the infrared image control points and the pixel coordinates of the visible light image control points to obtain the geographic coordinates: ； Where X represents the geographic horizontal coordinate, Y represents the geographic vertical coordinate, P represents the pixel column number, L represents the pixel row number, a represents the geographic horizontal coordinate corresponding to the top-left pixel, b represents the geographic resolution corresponding to the pixel width, c represents the row rotation parameter, d represents the geographic vertical coordinate corresponding to the top-left pixel, e represents the column rotation parameter, and f represents the geographic resolution corresponding to the pixel height. S304: Based on the geographic coordinates, perform a unified geographic coordinate system mapping on the pixel coordinates of the infrared image control points and the pixel coordinates of the visible light image control points: ； in, This represents the i-th ground control point. Represents the pixel coordinates in the infrared image. Represents the column coordinates in the infrared image. Represents the row coordinates in the infrared image. Represents latitude and longitude coordinates in the WGS84 coordinate system. Indicates latitude, Indicates longitude; S305: Based on the mapping results, generate a set of geographic control points containing the row and column numbers of control points, geographic coordinates, and image correspondences.

6. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, S4 specifically includes: S401: Based on the set of geographic control points, determine the matching point pairs between infrared image control points and reference image control points, and establish an initial affine transformation model between the matching point pairs; S402: Based on the initial affine transformation model, randomly select a preset number of matching point pairs from the matching point pairs, and calculate a temporary affine transformation model; S403: Based on the temporary affine transformation model, perform projection transformation on the remaining matching point pairs and calculate the projection error of each matching point pair; S404: Determine whether the projection error is less than the error threshold; if yes, determine the corresponding matching point pair as an interior point and add it to the current consistency control point set, and proceed to step S405; otherwise, determine the corresponding matching point pair as an exterior point and perform a removal mark. S405: Divide the target area into multiple regular grids, and assign each control point to the corresponding regular grid according to the geographical coordinates of each control point in the current set of consistent control points; S406: Calculate the spatial distribution characteristics of control points within each regular grid; S407: When the number of control points within the regular grid exceeds a preset threshold, calculate the comprehensive evaluation value based on the spatial distribution characteristics of the control points: ； Where p represents the control point to be evaluated. This represents the matching similarity weight coefficient. This represents the similarity score of the matching of control point p. This represents the current control point p to be evaluated and its neighboring control points. The distance between them; S408: Based on the comprehensive evaluation value, sort the control points within the regular grid, and determine the spatial homogenization control point set based on the sorting result; S409: The set of spatial homogenization control points is determined as the set of target control points.

7. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, S5 specifically includes: S501: Based on the geographical coordinates of each target control point in the target control point set, the target area is divided into regular grids to determine the large grid area and small grid area to which each target control point belongs; S502: Determine the corresponding control point type code, grid number, and year identifier based on the image data source, grid area, and image acquisition time of each target control point; S503: Generate the corresponding sequence number based on the generation order of each target control point within its respective small grid area; S504: Combine the control point type code, the grid number, the year identifier, and the serial number to generate a control point code; S505: Establish a spatial index relationship between the control point codes and the geographical locations of the target control points to obtain the set of coded control points.

8. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, S6 specifically includes: S601: Obtain the geographic coordinates and image pixel coordinates corresponding to each coded control point in the coded control point set; S602: Determine the center position of each coding control point based on the geographic coordinates and the image pixel coordinates; S603: Using the center position as the center, crop an infrared image window of a preset size from the infrared multispectral image to be matched, and crop a visible light image window of a preset size from the high-resolution visible light remote sensing image to obtain a dual-scale control point image block. S604: Based on the resolution ratio between the infrared multispectral image to be matched and the reference visible light image, a consistency constraint is imposed on the physical coverage of the infrared image window and the visible light image window. S605: Based on the consistency constraint results, determine whether the dual-scale control point image block exceeds the corresponding image boundary; if so, fill the region exceeding the boundary with background values; otherwise, retain the original image pixel values ​​and generate control point image blocks of uniform size. ； in, Indicates the cropped image block in Pixel value at the location, Indicates the location of the original image Pixel value at that location, This indicates that the current cropping position is located within the original image area. This indicates that the current cropping position is outside the original image area. This represents the background fill value.

9. The method for constructing an infrared remote sensing control point image database according to claim 1, characterized in that, The infrared remote sensing control point image database specifically includes: a physical storage layer, a spatial database management layer, a metadata organization layer, and a retrieval application layer.

10. A system for constructing an infrared remote sensing control point image database, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the infrared remote sensing control point image database construction method as described in any one of claims 1 to 9.

Images