Temporal and spatial fusion color correction method for remote sensing image based on virtual low-resolution reconstruction

Abstract

The application provides a kind of remote sensing image space-time fusion color correction method based on virtual low-resolution reconstruction, comprising: obtaining the original remote sensing image to be reconstructed, determining the corresponding reference image according to the characteristics of the original remote sensing image;Virtual low-resolution image reconstruction is carried out using the original remote sensing image and the reference image, and a virtual low-resolution image is obtained;Based on the virtual low-resolution image, the original remote sensing image is processed by color correction using a space-time fusion strategy to obtain a uniform color result image.In the color correction processing, a weighted space-time filtering mechanism is introduced, a virtual low-resolution image is generated as auxiliary data, and the time correlation, spatial proximity and spectral similarity between the original image and the reference image are considered comprehensively, the spectral similar pixels of the central pixel in the filtering window are used to participate in the result operation, and the result image with the original image detail characteristics and the reference image spectral characteristics is generated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and in particular to a remote sensing image spatio-temporal fusion color correction method based on virtual low resolution reconstruction. BACKGROUND

[0002] In remote sensing image processing, in order to meet the demand of large area remote sensing analysis, in practical application, multiple images need to be mosaicked to synthesize images covering a large area. However, due to the influence of seasons, light and atmospheric conditions and ground cover conditions, the colors of each image obtained in a large area are usually not uniform, which directly affects the interpretation and use of the mosaicked image. Therefore, in order to obtain high-quality large-area images with clear details and consistent colors, it is urgent to coordinate the color consistency between images before mosaicking, eliminate the color difference between different images, and provide protection for the accuracy of subsequent visual interpretation, production of digital orthophoto and other applications.

[0003] The existing color correction methods can be roughly divided into three categories according to different correction principles: histogram matching-based color matching method, statistical parameter-based color matching method and reference base map-based color matching method. The histogram matching-based method is simple in principle and easy to implement. The histogram of the original image is directly mapped to a specified shape according to certain rules to realize color adjustment. In the case that the overall color deviation of the image is not large and there is no sudden brightness distribution, the overall color tone between images can be coordinated. The statistical parameter-based color matching method believes that the statistical parameters representing color tone and brightness between images with balanced color distribution in a certain area should also have similarity. Therefore, the statistical parameters of multiple images in the area are transformed to make the colors consistent. The reference base map-based color matching method does not consider whether there is an overlapping area between adjacent images. It directly selects a low-resolution image covering the survey area and having uniform color from the existing historical images as a reference, and uses the corresponding relationship between the reference image and the original image to construct a model to process the original image. However, the histogram matching-based and statistical parameter-based color matching methods have certain limitations and are more effective between images with small areas and small color differences. Although the reference base map-based color matching method can effectively reduce error accumulation and human intervention, when using global statistical parameters or overall models to adjust and correct color differences, it is easy to cause loss of local detail information and degradation of quality characteristics of the image. Therefore, how to better balance color balance and detail clarity in the color matching process is a problem to be solved. SUMMARY

[0004] The present application provides a remote sensing image spatio-temporal fusion color correction method based on virtual low resolution reconstruction, which solves the defects in the prior art.

[0005] In a first aspect, the present application provides a remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction, comprising:

[0006] An original remote sensing image to be reconstructed is acquired, and a corresponding reference image is determined according to the characteristics of the original remote sensing image;

[0007] Virtual low-resolution image reconstruction is performed using the original remote sensing image and the reference image to obtain a virtual low-resolution image;

[0008] Based on the virtual low-resolution image, a spatio-temporal fusion strategy is used to perform color correction processing on the original remote sensing image to obtain a uniform color result image.

[0009] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the present application, an original remote sensing image to be reconstructed is acquired, and a corresponding reference image is determined according to the characteristics of the original remote sensing image, comprising:

[0010] Low-resolution images corresponding to the measurement area range and the corresponding time period of the original remote sensing image are downloaded from an open source network platform as the reference image;

[0011] Image data is obtained from historical processed achievement data of a data production unit as the reference image;

[0012] Color uniform images are selected from land satellite remote sensing images or pre-set wide images and processed as the reference image.

[0013] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the present application, virtual low-resolution image reconstruction is performed using the original remote sensing image and the reference image to obtain a virtual low-resolution image, comprising:

[0014] The mean and variance in the normal distribution of the original remote sensing image are determined respectively, and the mean and variance in the normal distribution of the reference image are determined;

[0015] Based on the mean and variance in the normal distribution of the original remote sensing image and the mean and variance in the normal distribution of the reference image, a histogram mapping relationship between the original remote sensing image gray value and the reference image gray value is established;

[0016] Color matching is performed using the histogram mapping relationship between the original remote sensing image gray value and the reference image gray value to obtain the virtual low-resolution image.

[0017] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the present application, virtual low-resolution image reconstruction is performed using the original remote sensing image and the reference image to obtain a virtual low-resolution image, further comprising:

[0018] determining the local mean and variance in the normal distribution of the original remote sensing image, and determining the local mean and variance in the normal distribution of the reference image;

[0019] determining the reference image gray value, and obtaining the virtual low-resolution image by using the local mean and variance in the normal distribution of the original remote sensing image, the local mean and variance in the normal distribution of the reference image, and the reference image gray value.

[0020] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the application, the virtual low-resolution image is reconstructed by using the original remote sensing image and the reference image, and the method further comprises:

[0021] extracting the background information of the reference image and the detail information of the original remote sensing image;

[0022] superimposing the background information and the detail information, replacing the background information by using the simulated background image in the color uniformization, and obtaining a color uniformization replacement image;

[0023] taking the mean value of the color uniformization replacement image as the center, combining a stretching coefficient to compensate for the brightness loss, and obtaining the virtual low-resolution image.

[0024] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the application, based on the virtual low-resolution image, the original remote sensing image is color corrected by using a spatio-temporal fusion strategy, and a color uniformization result image is obtained, comprising:

[0025] determining any candidate pixel as a similar pixel in the original remote sensing image based on a preset screening condition;

[0026] based on the similar pixel, calculating the spectral difference measure, the temporal difference measure and the spatial difference measure between the original remote sensing image, the reference image and the virtual low-resolution image;

[0027] using a weighted spatio-temporal filtering color correction to process the spectral difference measure, the temporal difference measure and the spatial difference measure, and outputting the color uniformization result image.

[0028] According to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction provided by the application, based on a preset screening condition, any candidate pixel in the original remote sensing image is determined as a similar pixel, comprising:

[0029] determining the threshold for screening the spectral similar pixel from the high-resolution image mean square deviation and the estimated ground object category number;

[0030] A center pixel is determined according to a local moving window size, and the absolute value of the difference between any candidate pixel in the original remote sensing image and the center pixel is less than a threshold value of the screened similar pixel, so as to obtain the similar pixel.

[0031] According to the present application, a remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction is provided, and based on the similar pixel, a spectral difference measure, a temporal difference measure and a spatial difference measure between the original remote sensing image, the reference image and the virtual low-resolution image are calculated, including:

[0032] The gray value difference of the original remote sensing image and the virtual low-resolution image at the same time is calculated, so as to obtain the spectral difference measure;

[0033] The gray value difference of the reference image and the virtual low-resolution image at different times is calculated, so as to obtain the temporal difference measure;

[0034] A specific constant is determined, and based on the specific constant, the spatial distance between the candidate pixel and the center pixel is calculated in a local moving window, so as to obtain the spatial difference measure.

[0035] According to the present application, a remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction is provided, and the spectral difference measure, the temporal difference measure and the spatial difference measure are processed by using a weighted spatio-temporal filtering color correction, and the color correction result image is output, including:

[0036] Based on the spectral difference measure, the temporal difference measure and the spatial difference measure, a weight function is calculated;

[0037] The color correction result image is calculated by the weight function and the gray values of the original remote sensing image, the reference image and the virtual low-resolution image at different times.

[0038] In a second aspect, the present application further provides a remote sensing image spatio-temporal fusion color correction system based on virtual low-resolution reconstruction, including:

[0039] An acquisition module is configured to acquire an original remote sensing image to be reconstructed, and determine a corresponding reference image according to the characteristics of the original remote sensing image;

[0040] A reconstruction module is configured to reconstruct a virtual low-resolution image by using the original remote sensing image and the reference image;

[0041] A correction module is configured to perform color correction processing on the original remote sensing image by using a spatio-temporal fusion strategy based on the virtual low-resolution image, so as to obtain a color correction result image.

[0042] Compared with the prior art, the present application has the following beneficial effects:

[0043] (1) The spatio-temporal fusion algorithm framework of remote sensing images is introduced into the color correction processing, and the present application proposes that the reference base map is reconstructed by using the spectral information of the original image to obtain a virtual low-resolution image as auxiliary data before color correction, the image is similar to the color distribution of the original image and close to the texture characteristics of the reference base map, which meets the requirements of the existing image in the fusion algorithm idea;

[0044] (2) The weighted spatio-temporal filtering is used to realize the color correction processing, the present application fully considers the time correlation, spatial proximity and spectral similarity of adjacent image elements between images, and the spectral similar image elements of the target image element are selected by using the local filtering window to reconstruct the uniform color image, so that the result image can retain the detail characteristics of the high-resolution original image to the maximum extent while keeping the color distribution close to the reference base map, and the deficiencies of the existing methods in considering the image details and the relationship between image elements can be effectively overcome. Figure One BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0046] Figure 1 is a process schematic diagram of the spatio-temporal fusion color correction of remote sensing images based on virtual low-resolution reconstruction provided by the present application;

[0047] Figure 2 is a principle diagram of the spatio-temporal fusion color correction of remote sensing images based on virtual low-resolution reconstruction provided by the present application;

[0048] Figure 3 is a structure schematic diagram of the spatio-temporal fusion color correction system of remote sensing images based on virtual low-resolution reconstruction provided by the present application;

[0049] Figure 4 is a structure schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0050] ​In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0051] In view of the problems in the prior art, the present application provides a remote sensing image space-time fusion color correction method based on virtual low-resolution reconstruction, a weighted space-time filtering mechanism is introduced in the color correction process, a virtual low-resolution image is generated as auxiliary data, and the time correlation, spatial proximity and spectral similarity among the virtual low-resolution image, the original image and the reference image are comprehensively considered, the spectral similar pixels of the central pixel in the filtering window are used to participate in the result operation, and a result image with original image detail features and reference image spectral features is generated.

[0052] Figure 1 It is a process schematic diagram of the remote sensing image space-time fusion color correction based on virtual low-resolution reconstruction provided by the embodiment of the present application, as shown in Figure 1 , comprising:

[0053] Step 100: acquiring an original remote sensing image to be reconstructed, and determining a corresponding reference image according to the characteristics of the original remote sensing image;

[0054] Step 200: reconstructing a virtual low-resolution image by using the original remote sensing image and the reference image;

[0055] Step 300: performing color correction processing on the original remote sensing image based on the virtual low-resolution image by using a space-time fusion strategy, and obtaining a uniform color result image.

[0056] Specifically, in the embodiment of the present application, first, in order to fuse the space-time information of images with different resolutions, the spectral information of the original image is used to reconstruct the reference image, and a virtual low-resolution image is obtained as auxiliary data before correction; second, in order to maximize the preservation of detail information, a weighted space-time filtering mechanism is introduced, the time correlation, spatial proximity and spectral similarity among the reconstructed image, the original image and the reference image are highlighted and quantified, so that they participate in the calculation of the correction result, and a result image with original image detail features and reference image spectral features is generated.

[0057] The technical solutions of the present application will be described in combination with Figure 2 .

[0058] First, the corresponding reference image is selected according to the image characteristics.

[0059] At present, remote sensing image data is large and comes from a wide range of sources, and it is usually necessary to obtain a reference base map with uniform color and convenient processing to carry out color matching, that is, historical achievement data with good color consistency and no splicing seam can be used as a reference base map.

[0060] The main ways to obtain a reference base map are as follows: first, low-resolution images corresponding to the survey area range and time period can be downloaded from open network platforms such as Google Maps, Tianditu and Baidu Maps as a reference base map; second, images can be obtained from historical achievement data processed by data production units; and third, images with uniform color can be selected from land satellite remote sensing images or images with a larger width, and then processed as a reference base map.

[0061] It can be understood that the spectral characteristics of different images covering the same area should be the same, so the geographical position is one of the factors to be considered when selecting a reference base map, that is, an image covering the same area as the original reference image is selected. Secondly, complex ground cover is usually changed at different times, and the imaging season should be similar to the original image when selecting a reference base map, so as to ensure that the ground cover type and distribution of the same area are similar. In addition, since the ultimate goal of color matching processing is to obtain an image with uniform color and rich details, the balance of color distribution of the image must be considered as an important standard. Therefore, the reference base map selected before color matching of remote sensing images based on the reference base map should have the characteristics of uniform color, covering the same survey area as the original image, similar imaging time and less ground cover change.

[0062] Secondly, after obtaining the corresponding reference image, a virtual low-resolution image is reconstructed using the reference image and the original image.

[0063] In fact, the reconstruction process of the virtual low-resolution image is to transfer information between the original image and the reference base map according to certain matching rules, so that the reconstructed image superimposes the color distribution consistent with the original image while maintaining the texture of the original reference base map. According to the characteristics of the virtual low-resolution image, the color information of the original image is extracted and matched to the reference base map. The methods based on histogram matching, statistical feature parameter matching and linear transformation can be used. First, the cumulative histogram, mean and variance of the original known image are recorded, and then the mapping relationship between them is established to realize color matching. Finally, a virtual low-resolution image with similar color distribution to the original image and texture characteristics of the reference base map is generated, and its quality is analyzed and evaluated. The image meeting the conditions is used as auxiliary data for constructing a color matching model for subsequent processing. Figure One

[0064] (1) The reconstruction method based on histogram matching takes the histogram of the original image to be color matched as the standard for the reference base map .​ histogram matching, the resulting histogram is consistent with the reference histogram, and the two images have good color consistency. Assuming that the distribution of , is normal, that the distribution of , is normal, then we have:

[0065]

[0066] Thus we can get:

[0067]

[0068] According to the above formula, the mapping relationship between the original image gray value and the reference base map gray value is established for color matching, and the virtual low-resolution image is obtained.

[0069] (2) The reconstruction method based on statistical characteristic parameter matching uses parameters representing image gray distribution and dispersion degree, such as mean value, variance and other statistical parameters, to establish the color relationship between the image to be processed and the reference standard to achieve color matching effect, among which, Wallis filtering is one of the most commonly used methods, and the conversion relationship can be expressed as:

[0070]

[0071] and are the gray values of the color matching reference base map and the reconstructed virtual low-resolution image respectively, and represent the local mean and variance of the original image to be color matched respectively, and represent the local mean and variance of the color matching reference base map .

[0072] (3) The linear transformation reconstruction method uses the linear correlation of the overlapping area between the image and the image to perform matching and interpolate it to the remaining part of the original image data set. It regards the image F as the superposition of the image detail information and the background image B:

[0073]

[0074] When color matching, the simulated background image is used to replace the background in the original image:

[0075]

[0076] Then the mean value of generating a spectral reconstruction image to compensate for the loss of brightness around the center :

[0077]

[0078] wherein, and is a stretching coefficient.

[0079] Thirdly, after obtaining the reconstructed virtual low-resolution image, the original image is color corrected by using a space-time fusion strategy.

[0080] (1) Find the similar pixels of the center pixel in the original image.

[0081] Without considering the influence of geometric registration and atmospheric correction errors between images, the pixels of multi-source images in the same region have correlation, and the homogeneous pixels of high resolution in the region can be calculated by using the adjacent low-resolution spectral similar pixels. In order to ensure the use of correct adjacent pixel information, only the pixel data with the same spectral type as the original image to be processed and without cloud and fog shielding can be used for calculation. The present application defines the pixels with a gray value difference less than a certain threshold in the neighborhood as similar pixels, and selects similar adjacent pixels to the center pixel in the image by using a local moving window to reconstruct the image. It is worth noting that too large window will introduce redundant mixed pixels, so the window size can be set to a suitable value according to the characteristics of the region covered by different images. Generally, the similar pixels are determined according to the following formula:

[0082]

[0083] wherein, is the size of the local moving window, which is determined according to the complexity of the image ground objects in the study area and experimental experience value; ( , ) and ( , ) are the candidate pixels and the center pixel respectively, and the constant is the threshold for screening spectral similar pixels, which can be determined by the mean square deviation of the high-resolution image and the estimated number of ground object categories , that is, .

[0084] (2) Quantify the spectral, temporal and spatial differences between the original image, the reference image and the low-resolution reconstructed image.

[0085] The reconstructed virtual image as the original image The low-resolution image at the corresponding same moment, between which there is a reflectivity difference between the high / low-resolution images at the moment, and the low-resolution reference base map at another moment There is a difference between the low-resolution images caused by different imaging times. Considering the complex land cover types and changing environmental effects, the present application quantifies the spectral, temporal and spatial differences between images on the basis of obtaining the reconstructed image, in order to reduce the errors caused by external conditions and achieve the purpose of considering the color differences between land cover categories.

[0086] The spectral difference between the original image to be homogenized and the virtual low-resolution image is represented as

[0087]

[0088] The temporal difference between the reference base map and the virtual low-resolution image is represented as

[0089]

[0090] , Indicates different moments of the image.

[0091] The spatial distance between the candidate pixel in the moving window and the center pixel is , and the calculation formula is as follows:

[0092]

[0093] Wherein, is a specific constant, and the relative distance in the moving window changes from 1 to , The smaller the value, the greater the dynamic range. , and The smaller the value, the smaller the difference between the adjacent pixel and the center pixel.

[0094] (3) Color correction processing is realized by using weighted filtering.

[0095] In order to apply the quantized spatio-temporal spectral difference to the homogenization processing, the present application introduces a weight function to participate in the homogenization fusion, which is determined by the spectral difference, temporal difference and spatial difference between the original image and the virtual low-resolution image and the reference base map, and comprehensively measures the importance of the similar pixels around the target pixel to its spectral information, and finally generates a homogenization result image by traversing the entire image.

[0096] The comprehensive quantized spatio-temporal spectral difference, the weight function​​ The normalization can be expressed as:

[0097]

[0098] represents the total number of similar pixels;

[0099] The calculation formula of the center pixel after local window filtering is obtained in combination with the above formula:

[0100]

[0101] represents different moments of the image.

[0102] As shown in Figure 2 The remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction includes two main parts: virtual low-resolution image reconstruction and spatio-temporal fusion color correction based on weighted spatio-temporal filtering. In order to combine high-resolution detail features and low-resolution spectral features, and effectively utilize the relationship between adjacent pixels, the method generates a virtual low-resolution reconstructed image as auxiliary data, introduces a weighted spatio-temporal filtering mechanism in the processing process, considers the time correlation, spatial proximity and spectral similarity between the center element of the filtering window and the pixels of the original image or the reference image, uses the spectral similar pixels of the center element of the filtering window to participate in the result operation, and finally generates a result image with original image detail features and reference image spectral features.

[0103] The remote sensing image spatio-temporal fusion color correction system based on virtual low-resolution reconstruction provided by the present application is described below. The remote sensing image spatio-temporal fusion color correction system based on virtual low-resolution reconstruction described below can be correspondingly referred to the remote sensing image spatio-temporal fusion color correction method based on virtual low-resolution reconstruction described above.

[0104] Figure 3 is a structural schematic diagram of the remote sensing image spatio-temporal fusion color correction system based on virtual low-resolution reconstruction provided by the embodiment of the present application, as shown in Figure 3 includes an acquisition module 31, a reconstruction module 32 and a correction module 33, wherein:

[0105] The acquisition module 31 is used to acquire an original remote sensing image to be reconstructed, and determine a corresponding reference image according to the characteristics of the original remote sensing image. The reconstruction module 32 is used to reconstruct a virtual low-resolution image by using the original remote sensing image and the reference image. The correction module 33 is used to perform color correction processing on the original remote sensing image based on the virtual low-resolution image by using a spatio-temporal fusion strategy, and obtain a color-uniform result image.

[0106] ​Figure 4 An example of a schematic diagram of a physical structure of an electronic device is shown in Figure 4 As shown, the electronic device can include a processor 410, a communications interface 420, a memory 430, and a communications bus 440, wherein the processor 410, the communications interface 420, and the memory 430 can communicate with each other through the communications bus 440. The processor 410 can invoke a logical instruction in the memory 430 to execute a virtual low-resolution reconstruction-based spatio-temporal fusion color correction method for remote sensing images, which includes: obtaining an original remote sensing image to be reconstructed, determining a corresponding reference image according to a feature of the original remote sensing image; performing virtual low-resolution image reconstruction using the original remote sensing image and the reference image to obtain a virtual low-resolution image; and performing color correction processing on the original remote sensing image based on the virtual low-resolution image using a spatio-temporal fusion strategy to obtain a uniform color result image.

[0107] In addition, the logical instruction in the memory 430 described above can be implemented in the form of a software functional unit and sold or used as an independent product, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0108] The device embodiments described above are only schematic, wherein the units shown as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment. Those skilled in the art can understand and implement it without creative labor.

[0109] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0110] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction, characterized in that: include: Acquire an original remote sensing image to be reconstructed, and determine a corresponding reference image based on the characteristics of the original remote sensing image; Reconstructing a virtual low-resolution image using the original remote sensing image and the reference image to obtain a virtual low-resolution image; Based on the virtual low-resolution image, a spatiotemporal fusion strategy is used to perform color correction processing on the original remote sensing image to obtain a uniform color result image, including: Determining any candidate pixel as a similar pixel in the original remote sensing image based on a preset screening condition; Based on the similar pixels, calculating the spectral difference metric, the temporal difference metric and the spatial difference metric among the original remote sensing image, the reference image and the virtual low-resolution image; Processing the spectral difference metric, the phase difference metric, and the spatial difference metric using weighted spatiotemporal filtering color correction to output the uniform color result image; Calculating a spectral difference metric, a temporal difference metric, and a spatial difference metric among the original remote sensing image, the reference image, and the virtual low-resolution image based on the similar pixels includes: Calculating the grayscale value difference between the original remote sensing image and the virtual low-resolution image at the same time to obtain the spectral difference measurement; Calculating the grayscale value difference between the reference image and the virtual low-resolution image at different times to obtain the phase difference metric; A specific constant is determined, and a spatial distance between the candidate pixel and the center pixel is calculated in a local moving window based on the specific constant to obtain the spatial difference metric.

2. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1 is characterized in that: Obtaining an original remote sensing image to be reconstructed, and determining a corresponding reference image based on the characteristics of the original remote sensing image, including: Downloading a low-resolution image with a corresponding measurement area and time period as the original remote sensing image from an open source network platform as the reference image; Acquire an image from historical achievement data processed by a data production unit as the reference image; An image with uniform color is selected from land satellite remote sensing images or images with preset width and processed as the reference image.

3. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1 is characterized in that: Reconstructing a virtual low-resolution image using the original remote sensing image and the reference image to obtain a virtual low-resolution image includes: respectively determining the mean and variance in the normal distribution of the original remote sensing image and determining the mean and variance in the normal distribution of the reference image; Establishing a histogram mapping relationship between the grayscale value of the original remote sensing image and the grayscale value of the reference image based on the mean and variance in the normal distribution of the original remote sensing image and the mean and variance in the normal distribution of the reference image; Color matching is performed using a histogram mapping relationship between the grayscale values ​​of the original remote sensing image and the grayscale values ​​of the reference image to obtain the virtual low-resolution image.

4. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1, characterized in that: Reconstructing a virtual low-resolution image using the original remote sensing image and the reference image to obtain a virtual low-resolution image also includes: respectively determining a local mean and a variance in a normal distribution of the original remote sensing image, and determining a local mean and a variance in a normal distribution of the reference image; The grayscale value of the reference image is determined, and the virtual low-resolution image is obtained by using the local mean and variance in the normal distribution of the original remote sensing image, the local mean and variance in the normal distribution of the reference image, and the grayscale value of the reference image.

5. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1 is characterized in that: Reconstructing a virtual low-resolution image using the original remote sensing image and the reference image to obtain a virtual low-resolution image also includes: extracting background information of the reference image and detail information of the original remote sensing image; The background information and the detail information are superimposed, and when performing color uniformity, a simulated background image is used to replace the background information to obtain a color uniformity replacement image; The virtual low-resolution image is obtained by taking the mean value of the uniform color replacement image as the center and combining the stretching coefficient to compensate for brightness loss.

6. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1, characterized in that: Determining any candidate pixel as a similar pixel in the original remote sensing image based on preset screening conditions includes: The threshold for screening spectrally similar pixels is determined by the mean square error of high-resolution images and the estimated number of ground object categories. The central pixel is determined according to the size of the local moving window, and the similar pixel is obtained when the absolute value of the difference between any candidate pixel in the original remote sensing image and the central pixel is less than the threshold value for screening spectrally similar pixels.

7. The remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to claim 1, characterized in that: The method of processing the spectral difference metric, the phase difference metric, and the spatial difference metric using weighted spatiotemporal filtering color correction to output the uniform color result image comprises: Calculating a weight function based on the spectral difference metric, the temporal phase difference metric, and the spatial difference metric; The uniform color result image is calculated based on the weight function and the grayscale values ​​of the original remote sensing image, the reference image and the virtual low-resolution image at different times.

8. A remote sensing image spatiotemporal fusion color correction system based on virtual low-resolution reconstruction, based on the remote sensing image spatiotemporal fusion color correction method based on virtual low-resolution reconstruction according to any one of claims 1 to 7, characterized in that: include: An acquisition module is used to acquire an original remote sensing image to be reconstructed and determine a corresponding reference image according to the characteristics of the original remote sensing image; A reconstruction module, configured to reconstruct a virtual low-resolution image using the original remote sensing image and the reference image to obtain a virtual low-resolution image; The correction module is used to perform color correction processing on the original remote sensing image based on the virtual low-resolution image by adopting a spatiotemporal fusion strategy to obtain a uniform color result image.

Images