Registration method and device of remote sensing image and electronic equipment

Abstract

The invention relates to the technical field of image processing, and provides a remote sensing image registration method and device and electronic equipment, and the method comprises the steps: selecting a target sampling rate from a plurality of reference image sampling rates, and carrying out the resampling of a down-sampled reference image, and obtaining a resampled reference image; carrying out feature point matching on the resampling reference image and the down-sampled remote sensing image to obtain a plurality of pairs of matching points; coordinates of the multiple pairs of matching points are adjusted to obtain multiple pairs of adjusted matching points; adjusting the resolution of the resampling reference image to be consistent with that of the image with unobvious features according to the mean value of the distance ratios of the multiple pairs of adjustment matching points to obtain a standard reference image; carrying out feature point matching on the image with unobvious features and the standard reference image; and according to a feature point matching result, carrying out image geometric correction on the image with unobvious features. Through the method, the registration precision of the remote sensing image with unobvious features is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a remote sensing image registration method, device and electronic equipment. Background Art

[0002] An image with unclear features refers to a remote sensing image in which the features of deserts, farmlands and other land objects are not obvious. In related technologies, the scale-invariant feature transform (SIFT) and its improved methods have rotation and scale invariance, and have been widely used in the automatic registration of remote sensing images, achieving good results. However, for images with unclear features, SIFT and its improved methods obtain relatively few matching points, or even no matching points, which cannot meet actual usage needs. Summary of the invention

[0003] The present invention provides a remote sensing image registration method, device and electronic equipment, which are used to solve the problem that the feature point matching process of an image with unclear features in the prior art obtains relatively few matching points or even no matching points, and improve the registration accuracy of the remote sensing image with unclear features.

[0004] The present invention provides a remote sensing image registration method, comprising: Downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; Selecting a target sampling rate from a plurality of benchmark image sampling rates, resampling the downsampled benchmark image to obtain a resampled benchmark image; and matching feature points of the resampled benchmark image with the downsampled remote sensing image to obtain a plurality of pairs of matching points; According to the sampling rate of the downsampling process and the sampling rate of the resampling process, the coordinates of the plurality of pairs of matching points are adjusted to obtain the plurality of pairs of adjusted matching points; According to the average of the distance ratios between the multiple pairs of adjusted matching points, the resolution of the resampled reference image is adjusted to be consistent with the image with unclear features, so as to obtain a standard reference image; The image with unclear features is matched with the standard reference image for feature points; and the image with unclear features is geometrically corrected according to the feature point matching result.

[0005] The remote sensing image registration method provided by the present invention further includes: Obtaining a resolution difference value between the resampled reference image and the downsampled remote sensing image; If the resolution difference value is greater than or equal to the resolution difference threshold, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

[0006] According to the remote sensing image registration method provided by the present invention, obtaining the resolution difference value between the resampled reference image and the downsampled remote sensing image includes: obtaining the resolution difference value according to the average of the distance ratios between multiple pairs of matching points.

[0007] The remote sensing image registration method provided by the present invention further includes: According to the preset sampling rate interval, a plurality of sampling rates are generated between the maximum sampling rate of the reference image and the minimum sampling rate of the reference image to obtain a plurality of reference image sampling rates.

[0008] According to the remote sensing image registration method provided by the present invention, the image with unclear features and the reference image are respectively downsampled to obtain the downsampled remote sensing image and the downsampled reference image, which specifically includes: Obtaining the image width and image height of the image with unclear features; Obtaining a sampling rate for downsampling processing according to the image width and the image height, and a preset downsampling image width and a preset downsampling image height; According to the sampling rate of the downsampling process, the image with unclear features and the reference image are downsampled respectively to obtain a downsampled remote sensing image and a downsampled reference image.

[0009] According to any one of the remote sensing image registration methods provided by the present invention, feature point matching is performed on the resampled reference image and the downsampled remote sensing image to obtain multiple pairs of matching points, including: Acquire a plurality of feature points of the downsampled remote sensing image; Acquire a preset range area centered on each feature point to obtain a template window corresponding to each feature point; and set the resampled reference image as a search window; According to the template window and the search window, grayscale template matching is performed on the resampled reference image and the downsampled remote sensing image to obtain multiple pairs of matching points.

[0010] The remote sensing image registration method provided by the present invention further includes: If the feature point matching fails, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

[0011] The present invention also provides a remote sensing image registration device, comprising: A downsampling processing module is used to downsample the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; A first matching module is used to select a target sampling rate from a plurality of reference image sampling rates, resample the downsampled reference image to obtain a resampled reference image; and perform feature point matching between the resampled reference image and the downsampled remote sensing image to obtain a plurality of pairs of matching points; A coordinate adjustment module, used to adjust the coordinates of multiple pairs of matching points according to the sampling rate of the downsampling process and the sampling rate of the resampling process to obtain multiple pairs of adjusted matching points; An image adjustment module, used to adjust the resolution of the resampled reference image to be consistent with the image with unclear features according to the average of the distance ratios between the multiple pairs of adjustment matching points, so as to obtain a standard reference image; The second matching module performs feature point matching on the image with unclear features and the standard reference image; and performs image geometry correction on the image with unclear features according to the feature point matching result.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the remote sensing image registration method as described in any one of the above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the remote sensing image registration method as described in any one of the above is implemented.

[0014] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the remote sensing image registration method as described in any one of the above is implemented.

[0015] The remote sensing image registration method, device and electronic device provided by the present invention downsample the image with unclear features and the reference image, so that in the subsequent process of selecting a target sampling rate from multiple reference image sampling rates to obtain a resampled reference image, the amount of data for feature point matching is reduced, thereby improving the feature point matching efficiency. Based on the sampling rates in the downsampling and resampling processes, multiple pairs of adjustment matching points are obtained to adjust the resolution of the reference image, and a standard reference image with the same resolution as the image with unclear features is obtained. In the case where the resolution of the image with unclear features is uncertain, a standard reference image that can achieve better feature point matching results is obtained for the image with unclear features. Not only is the number of matching points obtained in the feature point matching process of the image with unclear features increased, but also the registration accuracy of the image with unclear features is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is one of the flow charts of the remote sensing image registration method provided by the present invention.

[0018] Figure 2 This is the second flow chart of the remote sensing image registration method provided by the present invention.

[0019] Figure 3 This is the third flow chart of the remote sensing image registration method provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the remote sensing image registration device provided by the present invention.

[0021] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the descriptions used in this way can be interchanged where appropriate, so that the embodiments can be implemented in a sequence other than that illustrated or described in the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. The naming or numbering of the steps that appear in the present application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The process steps that have been named or numbered can change the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved. The division of modules in this application is a logical division. There may be other division methods when it is implemented in actual applications. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection between units can be electrical or other similar forms, which are not limited in this application. In addition, the modules or submodules described as separate components may or may not be physically separated, may or may not be physical units, or may be distributed in multiple circuit units, and some or all of the units may be selected according to actual needs to achieve the purpose of the present application.

[0024] The present invention provides a remote sensing image registration method, which is applied to a remote sensing image registration server to solve the problem in the prior art that a feature point matching process of an image with unclear features obtains relatively few matching points or even no matching points, and at the same time improves the registration accuracy of the remote sensing image with unclear features.

[0025] Combine the following Figure 1-Figure 5 The specific contents of the present invention are described.

[0026] Figure 1 This is one of the flow charts of a remote sensing image registration method provided by the present invention, including steps S101-S105.

[0027] Step S101 , downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image.

[0028] The image with unclear features in the embodiments of the present invention refers to a remote sensing image in which there are relatively large areas of deserts, farmlands and other areas with unclear features. "Relatively large" means that at least half of the areas in the remote sensing image are areas with unclear features. In this embodiment, the specific proportion of areas with unclear features in the selected image with unclear features is not specifically limited. It is understandable that other types of remote sensing images can also be set as images with unclear features for subsequent processing according to the actual needs of the user.

[0029] Specifically, in the subsequent steps, a target sampling rate may be selected from multiple baseline image sampling rates for multiple times to resample the downsampled baseline image. Therefore, repeatedly selecting the sampling rate and resampling the downsampled baseline image will increase the time for feature point matching. In order to improve the speed of feature point matching, it is necessary to downsample the image with unclear feature information and the baseline image according to the sampling rate of the downsampling process according to the size of the image with unclear feature information, so as to obtain a downsampled remote sensing image and a downsampled baseline image. This allows the subsequent feature point matching process for the image with unclear features to be performed on the downsampled remote sensing image.

[0030] In a possible implementation, in step S101, the image with unclear features and the reference image are respectively downsampled to obtain a downsampled remote sensing image and a downsampled reference image, which specifically includes steps S1011-S1013.

[0031] Step S1011, obtaining the image width and image height of the image with unclear features.

[0032] Step S1012, obtaining a sampling rate for downsampling processing according to the image width and the image height, and the preset downsampling image width and the preset downsampling image height.

[0033] Step S1013, downsampling the image with unclear features and the reference image respectively according to the sampling rate of the downsampling process to obtain a downsampled remote sensing image and a downsampled reference image.

[0034] Exemplarily, firstly, the preset downsampled image width or the preset downsampled image height of the downsampled remote sensing image is set to , the sampling rate during downsampling The calculation formula is as follows: ; in, max Indicates taking the maximum value, width and height are the image width and image height of the image with unclear features respectively.

[0035] when The value of , configure the preset downsampled image width to ,when The value of , configure the preset downsampled image height to . The value of can be set by the user according to the actual application process, and is not specifically limited in the embodiment of the present invention. S After the calculation is completed, use the sampling rate S The image with unclear features and the reference image are downsampled respectively to generate a downsampled remote sensing image and a downsampled reference image.

[0036] By adopting the above method, the reliability of downsampling can be improved in the process of downsampling the image with unclear features and the reference image. The noise and the influence of the unclear detail area in the image with unclear features are reduced by downsampling, and the resolution of the reference image is adjusted by resampling, so that the feature point matching is more accurate, thereby improving the number and quality of matching points.

[0037] Step S102, selecting a target sampling rate from a plurality of benchmark image sampling rates, resampling the downsampled benchmark image to obtain a resampled benchmark image; and matching feature points between the resampled benchmark image and the downsampled remote sensing image to obtain a plurality of pairs of matching points.

[0038] Specifically, the multiple benchmark image sampling rates include multiple sampling rates. A target sampling rate is selected from the multiple benchmark image sampling rates. The target sampling rate can be selected from the set starting from the minimum sampling rate among the multiple benchmark image sampling rates to obtain a resampled benchmark image. The resampled benchmark image is matched with the downsampled remote sensing image for feature points to obtain multiple pairs of matching points until the feature points are successfully matched.

[0039] In a possible implementation, Figure 2 As shown, in step S102, feature point matching is performed on the resampled reference image and the downsampled remote sensing image to obtain multiple pairs of matching points, including steps S201-S203.

[0040] Step S201, obtaining a plurality of feature points of a downsampled remote sensing image.

[0041] Specifically, feature points can be extracted from the downsampled remote sensing image by fast feature point extraction algorithms such as Harris and FAST. The present invention does not specifically limit the method used in the feature point extraction process of the downsampled remote sensing image, but only gives an example.

[0042] Step S202, obtaining a preset range area centered on each feature point, obtaining a template window corresponding to each feature point; and setting the resampled reference image as the search window.

[0043] Step S203, performing grayscale template matching on the resampled reference image and the downsampled remote sensing image according to the template window and the search window to obtain multiple pairs of matching points.

[0044] Specifically, taking each feature point as the center, obtain the surrounding M The preset range area of ​​M is used as the template window, and the resampled reference image is set as the search window. Through the sliding process of the template window in the search window, the similarity between the template window and the sub-image below the search window is calculated to find the matching position. Repeat the above operation for each feature point to obtain multiple pairs of matching points. The similarity between the template window and the sub-image below the search window is calculated by the image gray value.

[0045] By using grayscale template matching to obtain multiple pairs of matching points between the resampled reference image and the downsampled remote sensing image, the stability and matching success rate of the feature point matching process are improved.

[0046] In a possible implementation manner, in step S102, the process further includes: if the feature point matching fails, reselecting a target sampling rate from a plurality of reference image sampling rates, and performing resampling processing on the downsampled reference image.

[0047] The following is a specific example of a situation where feature point matching fails.

[0048] exist Figure 3 In the illustrated embodiment, it includes: steps S301 - S302 .

[0049] Step S301, obtaining a normalized correlation coefficient between a resampled reference image and a downsampled remote sensing image.

[0050] Step S302: if the normalized correlation coefficient is less than or equal to the normalized correlation coefficient threshold, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

[0051] Grayscale template matching mainly relies on the similarity of grayscale values, so it is sensitive to image rotation, scaling, and illumination changes. When there is a large rotation or scaling difference between the resampled reference image and the downsampled remote sensing image, grayscale template matching may not be able to accurately find the matching position, resulting in feature point matching failure.

[0052] Specifically, in the process of matching the feature points of the resampled reference image with the downsampled remote sensing image to obtain multiple pairs of matching points, if the normalized correlation coefficient between the resampled reference image and the downsampled remote sensing image is greater than the normalized correlation coefficient threshold, the feature point matching is considered successful. If the normalized correlation coefficient is less than or equal to the normalized correlation coefficient threshold, the feature point matching is considered to have failed, and it is necessary to reselect a target sampling rate from multiple reference image sampling rates, resample the downsampled reference image, and obtain a new resampled reference image. The new resampled reference image is then continued to be matched with the downsampled remote sensing image for feature points.

[0053] In a possible implementation, the following steps are also included.

[0054] According to the preset sampling rate interval, multiple sampling rates are generated between the maximum sampling rate of the reference image and the minimum sampling rate of the reference image to obtain multiple reference image sampling rates.

[0055] Specifically, the minimum sampling rate and the maximum sampling rate are determined according to the size of the reference image, and the size of the reference image after downsampling must be greater than the preset number of pixels, thereby improving the success rate of feature point matching. The maximum sampling rate is determined according to the scale difference between the image with unclear feature information and the reference image.

[0056] By adopting the above method, a reference image sampling rate set is constructed, and multiple sampling rates are generated according to a preset sampling rate interval, so that it can adapt to the geometric correction requirements of images with different resolutions, thereby enhancing the adaptability and flexibility of the image geometric correction process.

[0057] In a possible implementation manner, after obtaining multiple pairs of matching points in step S102, the following steps are further included: The resolution difference value between the resampled reference image and the downsampled remote sensing image is obtained; if the resolution difference value is greater than or equal to the resolution difference threshold, a target sampling rate is reselected from multiple reference image sampling rates to resample the downsampled reference image.

[0058] Specifically, the step of checking the feature point matching process is used to detect the feature point matching process of the grayscale template matching. Subsequently, after the feature point matching is completed, since the image feature information of the downsampled remote sensing image is not obvious, it may cause the grayscale template matching to be mismatched, and it is necessary to further determine whether the feature point is matched successfully.

[0059] By adopting the above method, the resolution difference value between the resampled reference image and the downsampled remote sensing image is obtained, and the sampling rate is reselected according to the value for resampling processing, so as to improve the registration accuracy of remote sensing images with unclear features.

[0060] In a possible implementation, obtaining a resolution difference value between a resampled reference image and a downsampled remote sensing image includes: The resolution difference value is obtained according to the average distance ratio between multiple pairs of matching points.

[0061] Specifically, because grayscale template matching itself is not scale-invariant, only when the resolutions of the two images are relatively consistent can they be correctly matched. Therefore, the distance ratio between multiple pairs of matching points can be used to determine whether the feature point matching is successful. The calculation formula is as follows: ; in, and For the For matching points and For matching points, , Represents the control points on the downsampled remote sensing image and The distance Represents the control points on the resampled reference image and Among them, the control points on the downsampled remote sensing image and the control points on the resampled reference image , forming the Matching points; Control points on downsampled remote sensing images and the control points on the resampled reference image , forming the Pair of matching points.

[0062] After all distance ratios are calculated, the mean of the distance ratios is calculated. If the mean is in the middle of the preset mean range, the feature point matching is considered successful and the iterative matching is stopped. Otherwise, the feature point matching is considered failed and the process returns to step S102 to reselect a target sampling rate from multiple benchmark image sampling rates and continue with subsequent steps.

[0063] Exemplarily, the preset mean value range may be configured as 0.8-1.2.

[0064] Step S103, adjusting the coordinates of multiple pairs of matching points according to the sampling rate of the downsampling process and the sampling rate of the resampling process to obtain multiple pairs of adjusted matching points.

[0065] Get the coordinates of the control points on the input downsampled remote sensing image and the resampled reference image. In the feature point matching process, the sampling rate is used for the remote sensing image. S The baseline image is first sampled using the sampling rate SThe samples are then sampled using the sampling rates of multiple benchmark image sampling rates. If the coordinates of the multiple pairs of matching points of the downsampled remote sensing image and the resampled benchmark image are and , then the coordinates of multiple pairs of adjusted matching points of the original input image with unclear features and the reference image are and The calculations are as follows: ; ; in, is the benchmark image sampling rate set SR ={ SR 1, SR 2, …, SR n} p Base image sampling rate.

[0066] Step S104, adjusting the resolution of the resampled reference image to be consistent with the image with unclear features according to the average of the distance ratios between the multiple pairs of adjusted matching points, to obtain a standard reference image.

[0067] Specifically, the above feature point matching is completed on the downsampled remote sensing image, resulting in fewer matching points and relatively low accuracy, so it is necessary to match the same-scale images. First, use multiple pairs of adjusted matching points and calculate the distance ratio mean according to the distance ratio calculation method in the above steps. The mean is the resolution difference between the image with unclear features and the reference image. Then use the distance ratio mean as the sampling rate to sample the reference image to obtain a standard reference image so that it has the same resolution as the image with unclear features.

[0068] Step S105, performing feature point matching on the image with unclear features and the standard reference image; and performing image geometry correction on the image with unclear features according to the feature point matching result.

[0069] Specifically, by matching feature points of an image with unclear features with a standard reference image, more pairs of matching points with higher accuracy can be obtained. These matching points are used to construct a correction model, thereby completing the geometric correction of the image with unclear features.

[0070] In the embodiment of the present invention, the specific manner of performing image geometric correction on the image with unclear features according to the feature point matching result is not specifically limited.

[0071] In a possible implementation manner, after the feature point matching performed in any step (including step S102 and step S105) obtains multiple pairs of matching points, the following further includes: If the number of matching points obtained by feature point matching is less than the first preset number of matching points, it is determined that this feature point matching has failed. If the number of control points is greater than the first preset number of matching points, RANSAC and other methods are used to remove the wrong matching points in multiple pairs of matching points. After the removal is completed, if the number of matching points is less than the second preset number of matching points, it is determined that this feature point matching has failed. After the feature point matching fails, it is necessary to reselect a target sampling rate from multiple reference image sampling rates, resample the downsampled reference image, and obtain a new resampled reference image.

[0072] By adopting the above method, the present invention can achieve the following beneficial effects: by downsampling the image with unclear features and the reference image, the amount of data for feature point matching is reduced in the process of selecting a target sampling rate from multiple reference image sampling rates to obtain a resampled reference image. Based on the sampling rates in the downsampling and resampling processes, multiple pairs of adjustment matching points are obtained to adjust the resolution of the reference image, and a standard reference image with the same resolution as the image with unclear features is obtained. In the case where the resolution of the image with unclear features is uncertain, a standard reference image that can achieve better feature point matching results is obtained for the image with unclear features. The number of matching points obtained in the feature point matching process of the image with unclear features is increased, and the registration accuracy of the remote sensing image with unclear features is improved.

[0073] Figure 4 A schematic diagram of the structure of a remote sensing image registration device provided by the present invention includes the following parts: A downsampling processing module 410 is used to downsample the image with unclear features and the reference image to obtain a downsampled remote sensing image and a downsampled reference image; The first matching module 420 is used to select a target sampling rate from multiple benchmark image sampling rates, resample the downsampled benchmark image to obtain a resampled benchmark image; and perform feature point matching between the resampled benchmark image and the downsampled remote sensing image to obtain multiple pairs of matching points.

[0074] The coordinate adjustment module 430 is used to adjust the coordinates of multiple pairs of matching points according to the sampling rate of the downsampling process and the sampling rate of the resampling process to obtain multiple pairs of adjusted matching points.

[0075] The image adjustment module 440 is used to adjust the resolution of the resampled reference image to be consistent with the image with unclear features according to the average of the distance ratios between the multiple pairs of adjustment matching points, so as to obtain a standard reference image.

[0076] The second matching module 450 performs feature point matching on the image with unclear features and the standard reference image; and performs image geometry correction on the image with unclear features according to the feature point matching result.

[0077] In a possible implementation, the device also includes a matching result detection module, which is used to obtain a resolution difference value between the resampled baseline image and the downsampled remote sensing image; if the resolution difference value is greater than or equal to a resolution difference threshold, a target sampling rate is reselected from multiple baseline image sampling rates to resample the downsampled baseline image.

[0078] In a possible implementation, the matching result detection module is specifically used to: The resolution difference value is obtained according to the average distance ratio between multiple pairs of matching points.

[0079] In a possible implementation, the device further includes a set construction module, which is used to construct a plurality of reference image sampling rates of the reference image, specifically used to: Configure the maximum and minimum sampling rates for the benchmark image; According to the preset sampling rate interval, multiple sampling rates are generated between the maximum sampling rate and the minimum sampling rate to obtain multiple reference image sampling rates.

[0080] In a possible implementation, the downsampling processing module 410 is specifically configured to: Get the image width and image height of the image with unclear features, Obtaining a sampling rate for downsampling processing according to the image width and the image height, as well as a preset downsampling image width and a preset downsampling image height; According to the sampling rate of the downsampling process, the image with unclear features and the reference image are downsampled respectively to obtain a downsampled remote sensing image and a downsampled reference image.

[0081] In a possible implementation manner, the first matching module 420 is specifically configured to: Obtain multiple feature points of downsampled remote sensing images; Taking each feature point as the center, obtain the area within the preset range to obtain the template window corresponding to each feature point; and set the resampled reference image as the search window; According to the template window and the search window, grayscale template matching is performed on the resampled reference image and the downsampled remote sensing image to obtain multiple pairs of matching points.

[0082] In a possible implementation manner, the first matching module 420 is further configured to: Obtaining the normalized correlation coefficient between the resampled reference image and the downsampled remote sensing image; If the normalized correlation coefficient is less than or equal to the normalized correlation coefficient threshold, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

[0083] Figure 5An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communications interface 520 and the memory 530 communicate with each other via the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute a remote sensing image registration method, the method comprising: Downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; Select a target sampling rate from multiple benchmark image sampling rates, resample the downsampled benchmark image to obtain a resampled benchmark image; and perform feature point matching between the resampled benchmark image and the downsampled remote sensing image to obtain multiple pairs of matching points; According to the sampling rate of the downsampling process and the sampling rate of the resampling process, the coordinates of the multiple pairs of matching points are adjusted to obtain multiple pairs of adjusted matching points; According to the average of the distance ratios between multiple pairs of adjusted matching points, the resolution of the resampled reference image is adjusted to be consistent with the image with unclear features, so as to obtain a standard reference image; The feature point matching is performed between the image with unclear features and the standard reference image; and the image geometric correction is performed on the image with unclear features according to the feature point matching result.

[0084] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0085] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute a remote sensing image registration method provided by the above methods, the method comprising: Downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; Select a target sampling rate from multiple benchmark image sampling rates, resample the downsampled benchmark image to obtain a resampled benchmark image; and perform feature point matching between the resampled benchmark image and the downsampled remote sensing image to obtain multiple pairs of matching points; According to the sampling rate of the downsampling process and the sampling rate of the resampling process, the coordinates of the multiple pairs of matching points are adjusted to obtain multiple pairs of adjusted matching points; According to the average of the distance ratios between multiple pairs of adjusted matching points, the resolution of the resampled reference image is adjusted to be consistent with the image with unclear features, so as to obtain a standard reference image; The feature point matching is performed between the image with unclear features and the standard reference image; and the image geometric correction is performed on the image with unclear features according to the feature point matching result.

[0086] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform a remote sensing image registration method provided by the above methods, the method comprising: Downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; Select a target sampling rate from multiple benchmark image sampling rates, resample the downsampled benchmark image to obtain a resampled benchmark image; and perform feature point matching between the resampled benchmark image and the downsampled remote sensing image to obtain multiple pairs of matching points; According to the sampling rate of the downsampling process and the sampling rate of the resampling process, the coordinates of the multiple pairs of matching points are adjusted to obtain multiple pairs of adjusted matching points; According to the average of the distance ratios between multiple pairs of adjusted matching points, the resolution of the resampled reference image is adjusted to be consistent with the image with unclear features, so as to obtain a standard reference image; The feature point matching is performed between the image with unclear features and the standard reference image; and the image geometric correction is performed on the image with unclear features according to the feature point matching result.

[0087] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0088] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A remote sensing image registration method, characterized in that: include: Downsampling the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; Selecting a target sampling rate from a plurality of benchmark image sampling rates, resampling the downsampled benchmark image to obtain a resampled benchmark image; and matching feature points of the resampled benchmark image with the downsampled remote sensing image to obtain a plurality of pairs of matching points; According to the sampling rate of the downsampling process and the sampling rate of the resampling process, the coordinates of the plurality of pairs of matching points are adjusted to obtain the plurality of pairs of adjusted matching points; According to the average of the distance ratios between the multiple pairs of adjusted matching points, the resolution of the resampled reference image is adjusted to be consistent with the image with unclear features, so as to obtain a standard reference image; Matching feature points of the image with unclear features and the standard reference image; And according to the feature point matching result, image geometric correction is performed on the image with unclear features.

2. The remote sensing image registration method according to claim 1, characterized in that: Also includes: Obtaining a resolution difference value between the resampled reference image and the downsampled remote sensing image; If the resolution difference value is greater than or equal to the resolution difference threshold, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

3. The remote sensing image registration method according to claim 2, characterized in that: The obtaining of the resolution difference value between the resampled reference image and the downsampled remote sensing image includes: obtaining the resolution difference value according to an average of distance ratios between multiple pairs of matching points.

4. The remote sensing image registration method according to claim 1, characterized in that: Also includes: According to the preset sampling rate interval, a plurality of sampling rates are generated between the maximum sampling rate of the reference image and the minimum sampling rate of the reference image to obtain a plurality of reference image sampling rates.

5. The remote sensing image registration method according to claim 1, characterized in that: The downsampling of the image with unclear features and the reference image to obtain the downsampled remote sensing image and the downsampled reference image specifically includes: Obtaining the image width and image height of the image with unclear features; Obtaining a sampling rate for downsampling processing according to the image width and the image height, and a preset downsampling image width and a preset downsampling image height; According to the sampling rate of the downsampling process, the image with unclear features and the reference image are downsampled respectively to obtain a downsampled remote sensing image and a downsampled reference image.

6. The remote sensing image registration method according to claim 1, characterized in that: The step of performing feature point matching on the resampled reference image and the downsampled remote sensing image to obtain a plurality of pairs of matching points includes: Acquire a plurality of feature points of the downsampled remote sensing image; Acquire a preset range area centered on each feature point to obtain a template window corresponding to each feature point; and set the resampled reference image as a search window; According to the template window and the search window, grayscale template matching is performed on the resampled reference image and the downsampled remote sensing image to obtain multiple pairs of matching points.

7. The remote sensing image registration method according to claim 6, characterized in that: Also includes: If the feature point matching fails, a target sampling rate is reselected from a plurality of reference image sampling rates to perform resampling processing on the downsampled reference image.

8. A remote sensing image registration device, characterized in that: include: A downsampling processing module is used to downsample the image with unclear features and the reference image respectively to obtain a downsampled remote sensing image and a downsampled reference image; A first matching module is used to select a target sampling rate from a plurality of reference image sampling rates, and resample the downsampled reference image to obtain a resampled reference image; and performing feature point matching between the resampled reference image and the downsampled remote sensing image to obtain a plurality of pairs of matching points; A coordinate adjustment module, used to adjust the coordinates of multiple pairs of matching points according to the sampling rate of the downsampling process and the sampling rate of the resampling process to obtain multiple pairs of adjusted matching points; An image adjustment module, configured to adjust the resolution of the resampled reference image to be consistent with the image with unclear features according to the average of the distance ratios between the multiple pairs of adjustment matching points, so as to obtain a standard reference image; A second matching module performs feature point matching between the image with unclear features and the standard reference image; And according to the feature point matching result, image geometric correction is performed on the image with unclear features.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the remote sensing image registration method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the remote sensing image registration method according to any one of claims 1 to 7 is implemented.

Images