An optimized image data acquisition method for national land space planning

Abstract

The present invention relates to the field of image enhancement technology, and specifically to an image data optimization acquisition method for land space planning. First, spectral reflectance information and mosaic images are obtained based on multiple remote sensing images of the planning area; then, the content similarity between different image blocks is analyzed while considering the detail expression degree of each image block in the entire mosaic result image, and the reference degree of the second image block as a reference image block of the first image block is obtained based on the quality evaluation and ground object similarity of the first image block and the second image block; finally, the uniform color result of all pixel points of the first image block is obtained based on the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained from each reference image block. The present invention can ensure the uniform color accuracy of remote sensing images after mosaicking, which is conducive to batch processing of image uniform color.

Description

Technical Field

[0001] The present invention relates to the field of image enhancement technology, and in particular to an image data optimization acquisition method for national land space planning. Background Art

[0002] National land space planning refers to a guiding plan formulated within a certain geographical area to rationally allocate and utilize land, resources and space based on comprehensive factors such as economy, society and environment. Images for national land space planning are usually remote sensing images collected by drones or remote sensing satellites. By extracting plots from remote sensing images, quantitative analysis of the spatial proportions of various types of land and objects can be achieved, and the data obtained from quantitative analysis can be used for national land space planning.

[0003] The remote sensing images used in national land space planning usually require large-area high-resolution remote sensing images that can cover the planned area. A single remote sensing image cannot meet this requirement, so it is necessary to mosaic and integrate multiple different remote sensing images into a larger complete image to meet the needs of national land space planning. The multiple remote sensing images used for image mosaicking may be obtained under different lighting conditions and different sensor conditions. Therefore, these remote sensing images need to be preprocessed when doing image mosaicking, such as geometric correction and registration operations. In addition, due to different lighting and other conditions, the tones of the images used for mosaicking will also be different. The tones will not only affect the mosaicking accuracy, but also may identify the same object as multiple different objects in the subsequent plot extraction due to the tones. At this time, directly enhancing the image blocks may cause some image content to be distorted, so the mosaicked images need to be processed for uniform color.

[0004] The patent with application number CN202111242549.8 discloses a large-area high-fidelity satellite remote sensing image color mosaic processing method, including screening fused images, and checking, supplementing and replacing fused images; mosaicking preprocessing the initial fused image, editing the mosaic lines of the preprocessed image, balancing the colors of the mosaic lines, and mosaicking, adjusting the clarity of the image, eliminating the influence of the atmosphere, refining the brightness and color of the local objects in the image, degrading the image, and outputting the image product. This method relies on the GXL platform and manual optimization to achieve image color uniformity. The color uniformity effect relies on the processing accuracy of the software platform and manual experience, and cannot guarantee the accuracy of image color uniformity, which is not conducive to batch processing of image color uniformity. Summary of the invention

[0005] In order to solve the above technical problems, the object of the present invention is to provide an image data optimization acquisition method for national land space planning, the method comprising:

[0006] Step S1: obtaining spectral reflectance information and mosaicking images based on multiple remote sensing images of the planning area;

[0007] Step S2: obtaining a quality evaluation of each image block according to the variance and range of the pixel values ​​of all pixels of each image block of the mosaicked image in each channel of the HSV color space;

[0008] Step S3: obtaining the ground object similarity between any two image blocks according to the difference in average emissivity of each distinguishing band in the spectral reflectance information of the two image blocks;

[0009] Step S4: obtaining a reference degree of the second image block as a reference image block of the first image block according to the quality evaluation and the ground object similarity of the first image block and the second image block;

[0010] Step S5: obtaining a uniform color result for all pixels of the first image block according to the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained from each reference image block.

[0011] Furthermore, step S1 specifically includes:

[0012] Obtain multiple remote sensing images of the area to be planned through the remote sensing satellite image library;

[0013] Acquiring spectral reflectance information of the ground areas corresponding to the multiple remote sensing images according to the ground object components of the ground areas corresponding to the multiple remote sensing images;

[0014] Perform image mosaicking on multiple remote sensing images to obtain mosaicked images.

[0015] Furthermore, step S2 specifically includes:

[0016] Obtaining a quality evaluation of the first image block according to the variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space and the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space;

[0017] Each image block of the mosaicked image is traversed to obtain a quality evaluation of each image block.

[0018] Furthermore, the variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space are positively correlated with the quality evaluation of the first image block; the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space is negatively correlated with the quality evaluation of the first image block.

[0019] Furthermore, step S3 specifically includes:

[0020] Dividing the spectrum into a plurality of bands at intervals of a preset first threshold of the spectrum length;

[0021] According to the distribution of spectral reflectance of each pixel point of each image block in each band, the discrimination degree of each band is obtained, and the multiple bands with the largest discrimination degrees are recorded as discrimination bands;

[0022] According to the discrimination degree of each discrimination band and the average emissivity difference of each discrimination band in the spectral reflectance information of any two image blocks, the ground object similarity of the two image blocks is obtained.

[0023] Furthermore, obtaining the discrimination of each band according to the distribution of the spectral reflectance of each pixel point of each image block in each band includes:

[0024] Cluster the reflectance of all pixels of each image block in a band to obtain multiple clusters;

[0025] The discrimination of the band is obtained according to the proportion of the corresponding pixel points in the largest cluster in the clustering result of the band among all the pixel points, the number of clusters in the clustering result of the band, and the extreme difference of the reflectance of all the pixel points in the band.

[0026] Furthermore, the proportion of the number of pixels corresponding to the largest cluster in the clustering result of the band among all the pixels is negatively correlated with the discrimination of the band, and the number of clusters in the clustering result of the band and the extreme difference of the reflectivity of all the pixels in the band are positively correlated with the discrimination of the band.

[0027] Furthermore, step S4 specifically includes: obtaining the reference degree of the second image block as a reference image block of the first image block according to the quality evaluation difference between the first image block and the second image block and the similarity between the first image block and the second image block, wherein the quality evaluation difference between the first image block and the second image block and the similarity between the first image block and the second image block are both positively correlated with the reference degree of the second image block as a reference image block of the first image block.

[0028] Furthermore, step S5 specifically comprises: calculating a weighted average of the mapping results of the first image block obtained from each reference image block to obtain a uniform color result for all pixels of the first image block, wherein the reference degree of each reference image block of the first image block is used as a weight.

[0029] Furthermore, after step S5, the method further includes: adjusting all pixel blocks of the mosaicked image according to the color uniformity result of each pixel point in each image block to obtain a final color uniformity result image.

[0030] The present invention has the following beneficial effects:

[0031] The present invention is aimed at remote sensing images for national land space planning. The present invention considers the detail expression degree of each image block in the entire mosaic result image and analyzes the content similarity between different image blocks, so as to obtain the quality evaluation of each image block and the ground object similarity between every two image blocks. According to the quality evaluation and ground object similarity of the first image block and the second image block, the reference degree of the second image block as the reference image block of the first image block is obtained; further, according to the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained from each reference image block, the uniform color result of all pixel points of the first image block is obtained; finally, the uniform color result of each pixel point in each image block is traversed, all pixel blocks of the mosaic image are adjusted, and the final uniform color result image is obtained, so as to ensure the accuracy of image uniform color and facilitate batch processing of image uniform color. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only 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.

[0033] Figure 1 An image that is not effectively color-optimized after mosaicking for national land space planning provided by an embodiment of the present invention;

[0034] Figure 2 A flow chart of an image data optimization acquisition method for national land space planning provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation method, structure, characteristics and effects of an image data optimization acquisition method for national land space planning proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. Unless otherwise defined, all technical and scientific terms used in this document have the same meaning as commonly understood by technicians in the technical field of the present invention.

[0036] The scenario targeted by the present invention is that when planning national land space, it is necessary to mosaic and integrate multiple different remote sensing images into a larger complete image to meet the needs of national land space planning. However, the multiple remote sensing images used for image mosaicking may be obtained under different lighting conditions and different sensor conditions. The mosaicked images may have different tones, resulting in areas where the image has obvious block phenomenon, such as the attached Figure 1 As shown, this type of image is not conducive to the use of national land space planning, so the mosaic image needs to be processed for uniform color to obtain a final image that can meet the needs of national land space planning.

[0037] The following is a detailed description of a specific scheme of an image data optimization collection method for national land space planning provided by the present invention in conjunction with the accompanying drawings.

[0038] See also Figure 2 , which shows a flow chart of an image data optimization acquisition method for national land space planning provided by an embodiment of the present invention, the method comprising the following steps:

[0039] Step S1: Obtain spectral reflectance information and mosaic images based on multiple remote sensing images of the planning area.

[0040] Wherein, step S1 specifically includes: obtaining multiple remote sensing images of the area to be planned through a remote sensing satellite image library;

[0041] According to the ground object components of the corresponding ground areas of the multiple remote sensing images, the spectral reflectance information of the corresponding ground areas of the multiple remote sensing images is obtained; and the multiple remote sensing images are mosaicked to obtain a mosaicked image.

[0042] More specifically, several remote sensing images of the area to be planned are obtained through the remote sensing satellite image library. The remote sensing image includes the spectral reflectance information of the corresponding ground area. One pixel point of the ground area image corresponds to a spectral reflectance curve, and the reflectance curve shows the reflectance of the ground object component corresponding to the pixel point to each band. According to the ground object components of the corresponding ground area of ​​multiple remote sensing images, the spectral reflectance information of the corresponding ground area of ​​multiple remote sensing images is obtained. Use the Seamless Mosaic tool in the software ENVI to mosaic several remote sensing images and obtain mosaicked images.

[0043] The mosaic image is composed of image blocks of multiple remote sensing images, in which the shooting quality of each image block is different. For example, the lighting conditions when some images are taken are dark, so the image blocks corresponding to the dark images will show low brightness and low contrast. When performing color grading, the contrast of each image block should be as high as possible, so that the details in the image block can be better expressed. However, for areas with low contrast, directly increasing the contrast will cause some details to be distorted. Therefore, the quality evaluation of each image block is first analyzed, and the quality evaluation reflects the contrast expression degree of each image block in the mosaic result image. The lower the quality evaluation, the lower the quality of the image block for the content detail texture in the image block, the greater the amplitude of the color adjustment required for the image block during color grading, and the lower the reference weight of the image block when used as a reference for other image blocks during color adjustment. Therefore, the present embodiment sets the following steps.

[0044] Step S2: Obtain the quality evaluation of each image block according to the variance and range of the pixel values ​​of all pixels of each image block of the mosaicked image in each channel of the HSV color space.

[0045] Among them, step S2 specifically includes: obtaining the quality evaluation of the first image block according to the variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space, and the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space; traversing each image block of the mosaicked image to obtain the quality evaluation of each image block.

[0046] More specifically, in order to analyze from a perspective more in line with the human visual system, the HSV (Hue, Saturation, Value) color space is selected for analysis. The range of hue is , the saturation range is , the brightness range is , normalize the hue value by dividing it by 360°, so that the hue value range is also within Each remote sensing image used for mosaicking in the mosaicking image corresponds to an image block, and the HSV pixel values ​​of all pixels of all image blocks in the mosaicking result image are obtained. The pixel value difference between each image block and the entire mosaicking result image is compared to obtain the quality evaluation of each image block.

[0047] In this embodiment, the calculation formula for the quality evaluation of each image block is constructed as follows:

[0048]

[0049] In the formula, Indicates the total number of channels. Here there are three channels: H, S, and V. Take A=3; Represents the variance of the pixel values ​​of all pixels in the i-th image block in the h-th channel; Represents the variance of the pixel values ​​of all pixels in the entire mosaic result image in the hth channel; Represents the extreme difference of the pixel values ​​of all pixels in the i-th image block in the h-th channel; represents linear normalization; Represents the quality evaluation of the i-th image block.

[0050] In the calculation formula for the quality evaluation of each image block constructed above, It represents the variance of the pixel value of the hth channel of the i-th image block, which reflects the distribution of the pixel values ​​of the i-th image block in the hth channel. The larger the variance, the greater the distribution difference of the pixel values. It represents the ratio of the variance of the pixel value of the hth channel of the ith image block to the variance of the pixel value of the hth channel of the entire image. This ratio reflects the richness of the distribution of the pixel values ​​of the hth channel compared with the entire image. reflects the contrast of the i-th image block, The larger the value of , the richer the pixel value distribution in the i-th image block and the greater the contrast. The more details the image block can show, and the greater the quality evaluation.

[0051] Therefore, the variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space are positively correlated with the quality evaluation of the first image block; the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space is negatively correlated with the quality evaluation of the first image block.

[0052] Similarly, the quality evaluation of all image blocks is obtained. The quality evaluation reflects the degree of expression of the content of the image block. The higher the quality evaluation, the better the expression quality of the content details and texture of the image block. The lower the quality evaluation, the lower the expression quality of the content details and texture of the image block. The greater the amplitude of the color adjustment required for the image block during the color uniformity processing, and the lower the reference weight of the image block when used as a reference for other image blocks during color adjustment.

[0053] The quality evaluation shows the degree of detail of each image block in the entire mosaic result image, but the content of each image block is not necessarily exactly the same. Some image blocks have relatively simple content and do not have too many details. The simple content leads to a low quality evaluation. Forcibly increasing the contrast of such simple image blocks may cause distortion of the content. Therefore, to evenly color the image, it is also necessary to analyze the content similarity between different image blocks, that is, the similarity of the types of objects. By comparing the content similarity between each image block and other image blocks, the colors of image blocks with similar content are adjusted to the same level, rather than directly increasing the contrast of all image blocks.

[0054] Since different ground objects on the ground have significantly different reflectivity to light in different bands, for example, green vegetation exhibits high reflectivity in the visible and near-infrared bands due to the strong absorption of chlorophyll, and soil generally has a high reflectivity in the visible to near-infrared bands, the similarity of the ground objects in two image blocks can be analyzed by the difference in the reflected light bands of the components in the image blocks. The more similar the ground objects are, the greater the reference degree to each other in color adjustment of the two image blocks. Therefore, this embodiment further sets the following steps.

[0055] Step S3: obtaining the ground object similarity between any two image blocks according to the difference in average emissivity of each distinguishing band in the spectral reflectance information of the two image blocks.

[0056] Wherein, step S3 specifically includes: dividing the spectrum into multiple bands with a preset spectral length first threshold as an interval; obtaining the discrimination of each band according to the distribution of spectral reflectance of each pixel point of each image block in each band, and recording the multiple bands with the largest discrimination as discrimination bands; obtaining the ground feature similarity of the two image blocks according to the discrimination of each discrimination band and the difference in average emissivity of each discrimination band in the spectral reflectance information of any two image blocks. Wherein, obtaining the discrimination of each band according to the distribution of spectral reflectance of each pixel point of each image block in each band includes: clustering the reflectance of all pixels points of each image block in a band to obtain multiple clusters; obtaining the discrimination of the band according to the proportion of the number of pixels corresponding to the largest cluster in the clustering result of the band in all pixels points, the number of clusters of the clustering result of the band, and the extreme difference of the reflectance of all pixels points in the band.

[0057] More specifically, when analyzing the similarity of objects between different image blocks through spectral analysis, since the spectral data has many bands and the data is complex, and the reflectivity differences of different objects in some bands are not large, first screening the bands with higher discrimination of different objects can reduce the amount of calculation and improve the operation speed, and also avoid excessive noise influence when too many bands are involved in the calculation. The discrimination of each band for different objects is measured by the reflectivity distribution of each pixel point of all image blocks in each band. The spectrum is divided into multiple bands with the preset first threshold of the spectrum length as the interval. In this embodiment, the first threshold of the spectrum length is set to 5nm, that is, the spectrum is divided into several bands with an interval of 5nm. Taking the tth band as an example, AP clustering is performed on the reflectivity of all pixels in the tth band to obtain several clusters. The calculation formula for constructing the discrimination of the tth band can be expressed as:

[0058]

[0059] In the formula, Indicates the ratio of the number of pixels corresponding to the largest cluster among all pixels in the clustering result of the reflectance of all pixels in the tth band; Indicates the number of clusters of the clustering result of the t-th band; Indicates the extreme difference of reflectivity of all pixels in the tth band; represents inverse proportional normalization; Indicates the discrimination of the t-th band.

[0060] In the calculation formula of the discrimination of the t-th band constructed above, It reflects the bias of the clustering result of the t-th band. The larger the value, the more pixels have similar reflectance in the t-th band, the lower the discrimination of the t-th band, and the number of clusters of the clustering result of the t-th band is. , the range of reflectance of all pixels in the tth band are positively correlated with the discrimination of the tth band. The larger the value of , the wider the distribution of the reflectance of each pixel point on the tth band, and the higher the discrimination of the tth band. Therefore, the proportion of the number of the largest cluster in all pixels in the clustering result of a band is negatively correlated with the discrimination of the band, and the number of clusters in the clustering result of the band and the extreme difference of the reflectance of all pixels in the band are positively correlated with the discrimination of the band.

[0061] Then, the discrimination degree of each band is obtained by the above method, and the multiple bands with the largest discrimination degree are recorded as discrimination bands. In this embodiment, the preset number of discrimination bands is K=10, and the k bands with the largest discrimination degree are obtained to compare the similarity between different image blocks, which are recorded as discrimination bands. According to the difference in reflectivity of different image blocks in these discrimination bands, the similarity of the ground objects between different image blocks is obtained. The more similar the ground objects are, the closer the reflectivity of the two image blocks in the discrimination bands is. In this embodiment, the calculation formula for constructing the ground object similarity between the i-th image block and the j-th image block can be expressed as:

[0062]

[0063] In the formula, Indicates the number of distinguished bands, here the preset number of distinguished bands is K=10; Indicates the discrimination of the kth discrimination band; represents the average reflectivity of all pixels in the i-th image block in the k-th distinguishing band, It represents the average reflectance of all pixels in the jth image block in the kth distinguishing band; Indicates the similarity between the i-th image block and the j-th image block; Denotes inverse normalization.

[0064] In the above-constructed calculation formula for the ground feature similarity between the i-th image block and the j-th image block, It represents the difference in average reflectance between the i-th image block and the j-th image block in the k-th distinguishing band. The average reflectance reflects the type of ground objects in the image block and the proportion of the ground objects in the image. The closer the difference in average reflectance between two image blocks is, the more consistent the ground object types of the two image blocks are and the closer the proportion is. As a weight, the greater the discrimination, the greater the weight of the difference in the band. It represents the weighted sum of the reflectivity differences of all distinguishing bands. Therefore, the distinction between any two distinguishing bands and the average emissivity difference of each distinguishing band in the spectral reflectivity information of any two image blocks are negatively correlated with the ground object similarity of the two image blocks. In this way, the ground object similarity between every two image blocks is obtained.

[0065] Step S4: according to the quality evaluation and the ground object similarity of the first image block and the second image block, obtain the reference degree of the second image block as a reference image block of the first image block.

[0066] Among them, step S4 specifically includes: according to the quality evaluation difference between the first image block and the second image block, and the similarity of the ground objects between the first image block and the second image block, obtaining the reference degree of the second image block as a reference image block of the first image block, and the quality evaluation difference between the first image block and the second image block and the similarity of the ground objects between the first image block and the second image block are both positively correlated with the reference degree of the second image block as a reference image block of the first image block.

[0067] More specifically, when the mosaicked image is color-leveled, the more similar the two image blocks are, the more similar the color range of the color-leveling results of the two image blocks should be. Through the histogram specification method (a well-known technology, which will not be described in detail in this embodiment), the histograms of each channel of the two similar image blocks are mapped to the same range so that the colors of the two image blocks are similar. The histogram specification method takes one image as the target and maps the grayscale range of the other image to the same range as the target image. In order to make the final color-leveling result more readable, the one with a higher quality evaluation of the two images is used as the target image.

[0068] First, obtain several reference image blocks for each image block. In this embodiment, define the first image block as the i-th image block, define the second image block as the j-th image block, and construct the reference degree calculation formula of the j-th image block as the reference image block of the i-th image block as follows:

[0069]

[0070] In the formula, represents the quality evaluation of the jth image block; Represents the quality evaluation of the i-th image block; Indicates the similarity between the i-th image block and the j-th image block; Indicates the reference degree of the j-th image block as the reference image block of the ith image block.

[0071] In the above-constructed reference degree calculation formula for taking the j-th image block as the reference image block of the i-th image block, Reflects the difference between the quality evaluation of the jth image block and the quality evaluation of the i-th image block. The value is positive when the quality evaluation of the j-th image block is greater, and the j-th image block is used as the reference image block of the i-th image block. Positive correlation; similarity between the i-th image block and the j-th image block , the larger its value is, the higher the reference degree of the j-th image block as the reference image block of the i-th image block is. Therefore, the reference degree of the j-th image block as the reference image block of the i-th image block is positive correlation; therefore, The larger the value of , the greater the quality evaluation of the j-th image block relative to the i-th image block and the more similar it is to the ground object of the i-th image block, and the greater the reference degree.

[0072] Step S5: obtaining a uniform color result for all pixels of the first image block according to the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained from each reference image block.

[0073] Step S5 specifically includes: calculating a weighted average of the mapping results of the first image block obtained from each reference image block to obtain a uniform color result for all pixels of the first image block, wherein the reference degree of each reference image block of the first image block is used as a weight.

[0074] More specifically, first, similarly to the method in step S4, the reference degrees of all image blocks as reference image blocks of the i-th image block can be obtained. The preset reference number is G=5, and the G image blocks with the largest reference degree and a reference degree greater than 0 are taken as the reference image blocks of the i-th image block. Here, only image blocks with larger reference images are taken. In order to be able to be used as target images for histogram specification in the future, when there are less than G blocks greater than 0, the maximum number that can be taken is taken. The reference degree sequence of the G reference image blocks of the i-th image block is normalized by the Norm function. In order to avoid the sequence being empty, the reference degree of the i-th image block to itself is Add to sequence , and record the normalized result as the normalized reference degree.

[0075] Then, obtain the histogram of all image blocks in each HSV channel. Take any channel as an example and take the histogram of the i-th image block as The reference image block is taken as the target image, and the histogram of the i-th image block is normalized to obtain the In this embodiment, the mathematical formula for constructing the uniform color result of each pixel in the i-th image block can be expressed as:

[0076]

[0077] In the formula, represents a reference number, in this embodiment, G=5; Indicates the length of the sequence after adding the reference degree of the i-th image block to itself; represents the first The reference degree of a reference image block is hour, Indicates the reference degree of the i-th image block to itself ; represents the mth pixel in the i-th image block. The mapping results; Represents the uniform color result of the m-th pixel in the ith image block.

[0078] In the mathematical formula for the uniform color result of each pixel in the i-th image block constructed above, represents the weighted mean of the mapping results of the i-th image block obtained from each reference image block, where the i-th image block The reference degree of the reference image block Represents weight.

[0079] Similarly, according to the above method, the uniform color results of all pixels in all image blocks are obtained.

[0080] Furthermore, after step S5, the method further includes: adjusting all pixel blocks of the mosaicked image according to the color uniformity result of each pixel point in each image block to obtain a final color uniformity result image.

[0081] Specifically, the color-homogenizing results of each channel of each pixel in all image blocks are used as the new pixel values ​​of each channel of each pixel. The pixels of all image blocks are adjusted to obtain the final color-homogenizing result image. The color-homogenizing result image is used for subsequent analysis such as plot extraction, and the analysis results are used for national land space planning.

[0082] In this embodiment, the quality evaluation of each image block is obtained according to the variance and range of the pixel values ​​of all pixels of each image block in each channel of the HSV color space of the mosaic image, and the ground feature similarity of any two image blocks is obtained by the average emissivity difference of each distinguishing band in the spectral reflectance information, and then the reference degree of the second image block as the reference image block of the first image block is obtained according to the quality evaluation and ground feature similarity of the first image block and the second image block; further, the uniform color result of all pixels of the first image block is obtained according to the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained by each reference image block, and finally, the uniform color result of each pixel in each image block is traversed, and all pixel blocks of the mosaic image are adjusted to obtain the final uniform color result image. The uniform color accuracy of the image used for national land space planning is guaranteed, which is conducive to batch processing of image uniform color.

[0083] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0084] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for optimizing image data collection for national land space planning, characterized in that: The method comprises: Step S1: obtaining spectral reflectance information and mosaicking images based on multiple remote sensing images of the planning area; Step S2: obtaining a quality evaluation of each image block according to the variance and range of the pixel values ​​of all pixels of each image block of the mosaicked image in each channel of the HSV color space; Step S3: obtaining the ground object similarity between any two image blocks according to the difference in average emissivity of each distinguishing band in the spectral reflectance information of the two image blocks; Step S4: obtaining a reference degree of the second image block as a reference image block of the first image block according to the quality evaluation and the ground object similarity of the first image block and the second image block; Step S5: obtaining a uniform color result for all pixels of the first image block according to the reference degree of each reference image block of the first image block and the mapping result of the first image block obtained from each reference image block; Step S3 specifically includes: Dividing the spectrum into a plurality of bands at intervals of a preset first threshold of the spectrum length; According to the distribution of spectral reflectance of each pixel point of each image block in each band, the discrimination degree of each band is obtained, and the multiple bands with the largest discrimination degrees are recorded as discrimination bands; According to the discrimination degree of each discrimination band and the average emissivity difference of each discrimination band in the spectral reflectance information of any two image blocks, the ground object similarity of the two image blocks is obtained; The obtaining of the discrimination of each band according to the distribution of the spectral reflectance of each pixel point of each image block in each band includes: Cluster the reflectance of all pixels of each image block in a band to obtain multiple clusters; The discrimination of the band is obtained according to the proportion of the corresponding pixel points in the largest cluster in the clustering result of the band among all the pixel points, the number of clusters in the clustering result of the band, and the extreme difference of the reflectance of all the pixel points in the band.

2. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: Step S1 specifically includes: Obtain multiple remote sensing images of the area to be planned through the remote sensing satellite image library; Acquiring spectral reflectance information of the ground areas corresponding to the multiple remote sensing images according to the ground object components of the ground areas corresponding to the multiple remote sensing images; Perform image mosaicking on multiple remote sensing images to obtain mosaicked images.

3. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: Step S2 specifically includes: Obtaining a quality evaluation of the first image block according to the variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space and the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space; Each image block of the mosaicked image is traversed to obtain a quality evaluation of each image block.

4. The image data optimization acquisition method for national land space planning according to claim 3 is characterized in that: The variance and range of the pixel values ​​of all pixels of the first image block of the mosaicked image in each channel of the HSV color space are positively correlated with the quality evaluation of the first image block; the variance of the pixel values ​​of all pixels of the mosaicked image in each channel of the HSV color space is negatively correlated with the quality evaluation of the first image block.

5. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: The proportion of the number of pixels corresponding to the largest cluster in the clustering result of the band among all the pixels is negatively correlated with the discrimination of the band, and the number of clusters in the clustering result of the band and the extreme difference of the reflectivity of all the pixels in the band are positively correlated with the discrimination of the band.

6. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: Step S4 specifically includes: obtaining the reference degree of the second image block as a reference image block of the first image block according to the quality evaluation difference between the first image block and the second image block and the similarity between the first image block and the second image block, wherein the quality evaluation difference between the first image block and the second image block and the similarity between the first image block and the second image block are both positively correlated with the reference degree of the second image block as a reference image block of the first image block.

7. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: Step S5 specifically includes: calculating a weighted average of the mapping results of the first image block obtained from each reference image block to obtain a uniform color result for all pixels of the first image block, wherein the reference degree of each reference image block of the first image block is used as a weight.

8. The image data optimization acquisition method for national land space planning according to claim 1 is characterized in that: The step S5 further includes: adjusting all pixel blocks of the mosaicked image according to the color uniformity result of each pixel point in each image block to obtain a final color uniformity result image.

Images