A Web map tile update processing method and device
By performing geometric correction, cloud shadow detection and splicing line replacement on remote sensing images, combined with Poisson fusion processing, the problem of poor stability and accuracy in web map tile updates is solved, and high-quality tile data updates are achieved.
Patent Information
- Application Number
- CN202111272026.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-29
- Publication Date
- 2024-12-13
- Estimated Expiration
- 2041-10-29
Smart Images
Figure CN114092326B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of Web map tile update processing, and in particular to a Web map tile update processing method and device. In addition, the present invention also relates to an electronic device and a processor-readable storage medium. Background Art
[0002] Web map tiles are map tile data that pre-cut remote sensing image data into square map data of fixed size (such as 256 pixels * 256 pixels) and specific format according to certain standards, and store them in the database according to specifications to form a pyramid model. Web map tiles adopt the strategy of "trading space for time" to speed up the server response speed. Therefore, when the client requests map data, the original remote sensing image data is no longer cut in real time, but the map slices are directly fed back to the client through the WMTS (OpenGIS Web Map Tile Service) slice map Web service. Web map tile update processing technology for Web map tiles has become a hot research topic under the urgent needs of engineering projects.
[0003] Web map tiles use tile matrix sets to represent the cut maps. "Tile" is an image of fixed size. A remote sensing image is cut into multiple tiles according to fixed rules to form a tile matrix. One or more tile matrices constitute a tile matrix set. Different tile matrices have different resolutions. Each tile matrix is identified by a tile matrix identifier (the serial number of the tile matrix, the layer with the lowest resolution is layer 0, and so on). The number of each tile needs to include layer Z, row number X, and column number Y. For the tile position of the specified layer, row number, and column number in the storage database, the Web map tile database usually has two storage methods. One is to store only one remote sensing map tile of the latest imaging, and the other is to store N remote sensing map tiles acquired at different times. For the latter tile storage database, the update is relatively simple. After the new remote sensing image data is tiled, it can be directly stored in the database after adding the imaging timestamp. However, the storage update of the former tile database, usually called the update of the basemap, is much more complicated and difficult. Since the basemap has higher requirements for geographic location accuracy, the absolute positioning accuracy of the objects in the map tiles is required to be as high as possible. After the basemap is updated with the newly acquired remote sensing image, it is required that the updated area has no obvious seams with the surrounding area, and the color and brightness are as consistent as possible with the surrounding area, forming a visual whole. This requires the remote sensing image to be processed by stitching line search, uniform light and color during the tile processing. At the same time, the basemap is required to be as cloudless as possible. According to statistics, the global average cloud coverage in the remote sensing images acquired by satellites exceeds 60%. This requires that the remote sensing images used for updating are clear and cloudless, or the clouds and shadows under the clouds can be marked during the update process, and the marked areas are filled with the original tile data when updating. Therefore, how to design an accurate and stable Web map tile update processing solution has become a problem that needs to be solved urgently. Summary of the invention
[0004] To this end, the present invention provides a Web map tile update processing method and device to solve the problem that the Web map tile update processing in the prior art has high limitations, resulting in poor stability and accuracy of the remote sensing image update reference base map.
[0005] In a first aspect, the present invention provides a method for processing a Web map tile update, comprising:
[0006] Acquire an input image; the input image is a remote sensing image to be tiled;
[0007] Determine a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data;
[0008] Performing geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy;
[0009] Performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determining a clear surface area of the standard input image based on the mask image;
[0010] Taking the reference image as a base image, determining a target stitching line between the input image and the reference image; replacing an excluded area in the input image with a corresponding area of the reference image based on the target stitching line to obtain a target input image;
[0011] The target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database.
[0012] Furthermore, before tiling the target input image, the method further includes:
[0013] The target input image and the reference image are input into a preset remote sensing image even-lighting and color-matching model based on Poisson fusion for even-lighting and color-matching processing to obtain the processed target input image; wherein the corresponding Poisson fusion area in the remote sensing image even-lighting and color-matching model is the clear surface area of the mask image of the standard input image, and the pixels of the reference image corresponding to the boundary of the clear surface area are used as guide vectors, and the guide vector domain is calculated toward the inside of the clear surface area to achieve even-lighting and color-matching processing.
[0014] Further, determining a reference image corresponding to the input image specifically includes:
[0015] Determining a maximum level of tiling according to a resolution of the input image after reprojection onto network map tiles; the resolution of the input image is not less than the determined maximum level;
[0016] Determining corresponding geographic coordinate ranges based on all network map tile images in the reference image that intersect with the geographic range of the input image;
[0017] Determine the update range of the network map tiles according to the geographic coordinate range, obtain the tile image of the maximum level within the update range, and stitch the tile image of the maximum level into a complete remote sensing image; and use the complete remote sensing image as a reference image corresponding to the input image.
[0018] Further, the input image is geometrically precisely corrected based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy, specifically including:
[0019] Extracting feature points from the input image and the reference image respectively, and performing feature matching on the feature points of the input image and the feature points in the reference image to obtain corresponding control point data;
[0020] A corresponding geometric correction model is determined based on the control point data, and the input image is input into the geometric correction model for geometric precision correction to obtain a standard input image that meets a preset geometric registration accuracy.
[0021] Furthermore, cloud and cloud shadow detection is performed on the standard input image to obtain a mask image of the standard input image, specifically including: inputting the standard input image and the reference image into a cloud and cloud shadow detection model for change detection, and using a guided filtering model to correct the boundary between clouds and shadows to obtain the mask image of the standard input image.
[0022] Further, taking the reference image as a base image, determining a target stitching line between the input image and the reference image specifically includes:
[0023] Using the reference image as the base image of the standard input image, determining a multi-band array composed of bands between the input image and the reference image based on a preset remote sensing image stitching line search model, and determining a target stitching line based on the multi-band array;
[0024] Replacing the excluded area in the input image with the corresponding area of the reference image based on the target stitching line specifically includes:
[0025] The target stitching line is determined as a boundary between a clear surface area and an exclusion area in the mask image; and based on the boundary, the exclusion area in the input image is replaced with a corresponding area of the reference image to obtain a target input image.
[0026] Furthermore, the target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database, specifically including: processing the target input image into tile image data in a preset format according to a preset tile division rule, and using the tile image data to update the original map tile data in the network map reference base map database.
[0027] In a second aspect, the present invention further provides a Web map tile update processing device, comprising:
[0028] A remote sensing image acquisition unit, used to acquire an input image; the input image is a remote sensing image to be tiled;
[0029] A reference image determination unit, used to determine a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data;
[0030] A geometric precision correction unit, configured to perform geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy;
[0031] a cloud and cloud shadow detection unit, configured to perform cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determine a clear surface area of the standard input image based on the mask image;
[0032] An image stitching line determining unit is used to determine a target stitching line between the input image and the reference image using the reference image as a base image; based on the target stitching line, an excluded area in the input image is replaced with a corresponding area of the reference image to obtain a target input image;
[0033] The tile processing unit is used to perform tile processing on the target input image and use the generated tile data to update the original map tile data in the network map reference base map database.
[0034] Furthermore, the Web map tile update processing device also includes: a uniform light and color processing unit, which is used to input the target input image and the reference image into a preset remote sensing image uniform light and color model based on Poisson fusion for uniform light and color processing before tiling the target input image, so as to obtain the processed target input image; wherein the corresponding Poisson fusion area in the remote sensing image uniform light and color model is the clear surface area of the mask image of the standard input image, and the pixels of the reference image corresponding to the boundary of the clear surface area are used as guide vectors, and the guide vector domain is calculated toward the inside of the clear surface area to achieve uniform light and color processing.
[0035] Furthermore, the reference image determination unit is specifically used to:
[0036] Determining a maximum level of tiling according to a resolution of the input image after reprojection onto network map tiles; the resolution of the input image is not less than the determined maximum level;
[0037] Determining corresponding geographic coordinate ranges based on all network map tile images in the reference image that intersect with the geographic range of the input image;
[0038] Determine the update range of the network map tiles according to the geographic coordinate range, obtain the tile image of the maximum level within the update range, and stitch the tile image of the maximum level into a complete remote sensing image; and use the complete remote sensing image as a reference image corresponding to the input image.
[0039] Furthermore, the geometric precision correction unit is specifically used for:
[0040] Extracting feature points from the input image and the reference image respectively, and performing feature matching on the feature points of the input image and the feature points in the reference image to obtain corresponding control point data;
[0041] A corresponding geometric correction model is determined based on the control point data, and the input image is input into the geometric correction model for geometric precision correction to obtain a standard input image that meets a preset geometric registration accuracy.
[0042] Furthermore, the cloud and cloud shadow detection unit is specifically used to: input the standard input image and the reference image into the cloud and cloud shadow detection model for change detection, and use the guided filtering model to correct the boundary between the cloud and the shadow to obtain the mask image of the standard input image.
[0043] Further, the image stitching line determination unit includes: a stitching line determination subunit and an exclusion area replacement subunit;
[0044] The stitching line determination subunit is specifically used to: use the reference image as the base image of the standard input image, determine a multi-band array composed of bands between the input image and the reference image based on a preset remote sensing image stitching line search model, and determine a target stitching line based on the multi-band array;
[0045] The exclusion area replacement subunit is specifically used to: determine the target stitching line as the boundary between the clear surface area and the exclusion area in the mask image; and replace the exclusion area in the input image with the corresponding area of the reference image based on the boundary to obtain the target input image.
[0046] Furthermore, the tile processing unit is specifically used to: process the target input image into tile image data of a preset format according to a preset tile division rule, and use the tile image data to update the original map tile data in the network map reference base map database.
[0047] In a third aspect, the present invention further provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, the steps of the Web map tile update processing method as described in any one of the above items are implemented.
[0048] In a fourth aspect, the present invention further provides a processor-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the Web map tile update processing method as described in any one of the above items.
[0049] The Web map tile update processing method provided by the embodiment of the present invention performs geometric precision correction on the input image through a reference image to obtain a standard input image that meets the preset geometric registration accuracy, and performs cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determines the clear surface area of the standard input image based on the mask image, and uses the reference image as the base map to determine the target stitching line between the input image and the reference image, and replaces the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain the target input image; the target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database. This method can make the generated map tile data have higher quality Web map tile update processing. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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.
[0051] Figure 1 A schematic diagram of a process flow of a Web map tile update processing method provided by an embodiment of the present invention;
[0052] Figure 2 A complete flowchart of a Web map tile update processing method provided by an embodiment of the present invention;
[0053] Figure 3 A schematic diagram of tile layer segmentation and numbering provided by an embodiment of the present invention;
[0054] Figure 4 A schematic diagram of the third layer of tile layering provided in an embodiment of the present invention;
[0055] Figure 5 Provided for the embodiment of the present invention is a schematic diagram of an input image and a reference image;
[0056] Figure 6 A schematic diagram of target splicing line search provided by an embodiment of the present invention;
[0057] Figure 7A schematic diagram of a Poisson fusion process provided by an embodiment of the present invention;
[0058] Figure 8 A schematic diagram of the structure of a Web map tile update processing device provided by an embodiment of the present invention;
[0059] Fig. 9 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0060] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments 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.
[0061] The purpose of the present invention is to provide a Web map tile update processing method for the preprocessing application of remote sensing image production network map tiles (i.e., Web map tiles), especially for the preprocessing application of the global base map Web map tiles. The present invention is based on remote sensing image geometric precision correction, remote sensing image cloud and cloud shadow detection algorithm, splicing line search algorithm, Poisson Blending algorithm and other technologies, and customizes the base map Web map tile data update method according to specific project applications. The remote sensing image to be tiled is used as the input image, and the reference image is downloaded from the web map benchmark base map database according to the input image. The geometric alignment accuracy of the input image and the reference image is made to reach the sub-pixel level by geometric precision correction to obtain the standard input image. The cloud and cloud shadow detection is performed on the standard input image to obtain a clear surface range image. The target stitching line between the clear surface range in the standard input image and the reference image is found to obtain the target input image. The target input image is subjected to uniform light and color processing to eliminate color and brightness differences. Finally, the processed high-quality target input image is tiled, and the generated tile data is used to update the original map tile data (i.e., the original benchmark base map Web map tile data) in the web map benchmark base map database (Web map benchmark base map database). There are no obvious "patches" and "seams" in the updated Web map, and the color and brightness of the updated area are also consistent with the original benchmark base map. Figure 1 To.
[0062] The following is a detailed description of an embodiment of the Web map tile update processing method according to the present invention. Figure 1As shown, it is a flow chart of a Web map tile update processing method provided by an embodiment of the present invention, and the specific implementation process includes the following steps:
[0063] Step 101: Acquire an input image; the input image is a remote sensing image to be tiled.
[0064] The remote sensing images to be tiled include multispectral and hyperspectral remote sensing images containing visible light RGB (RGB color mode) bands acquired by satellites, aerial photography or drones. The remote sensing images have map projection and coordinate information.
[0065] Step 102: Determine a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data.
[0066] In an embodiment of the present invention, the maximum level of tiling can be determined based on the resolution of the input image after reprojection to the web map tiles; and the corresponding geographic coordinate range can be determined based on all web map tile images in the reference image that intersect with the geographic range of the input image; the update range of the web map tiles is determined based on the geographic coordinate range, and the tile images of the maximum level within the update range are obtained, and the tile images of the maximum level are spliced into a complete remote sensing image; and the complete remote sensing image is used as the reference image corresponding to the input image.
[0067] like Figure 2 As shown, in an embodiment of the present invention, a Web (network) map tile database and a WMTS (WebMap Tile Service) service (i.e., Web map tile service) portion are also included. The Web map tiles can represent the cut map through a tile matrix set. A remote sensing image can be cut into multiple tiles according to a fixed size (such as 256*256 pixel size) to form a tile matrix. One or more tile matrices constitute a tile matrix set. For tile positions of specified levels, row numbers, and column numbers in the storage database, the Web map tile base map database only stores a set of tile data in the global format. An example diagram of tile layer segmentation and numbering is shown in FIG. Figure 3As shown, the global range is defined in tile layers, and each tile matrix is identified by a tile matrix identifier (the serial number of the tile matrix, the layer with the lowest resolution is layer 0, and so on), and the number of each tile needs to include layer Z, row number X, and column number Y. For example, the Zth layer of tiles, cutting to the Z+1th layer contains 4 tiles, and cutting to the Z+2th layer contains 16 tiles. Different tile matrix layers have different resolutions, and usually the 0th layer is a global map. For example, the resolution of WGS84 (World Geodetic System 1984) is 360 / 256 degrees in longitude and latitude, the resolution of the first layer is 180 / 256 degrees, the second layer is 90 / 256, and so on. Figure 4 This is an example of the third layer of tile stratification, and the global area is divided into 64 tiles of 256*256 pixels. Among them, the WMTS service can support publishing in HTTP KVP, SOAP and REST, and specifically includes operations such as obtaining metadata of the service, obtaining slices, and obtaining information of the selected elements. It can realize the reading of WMTS services and obtain tile images in jpg or png format. The Web map tile database is a device that actually stores tiles. The query and update of tile data can be completed through the WMTS service link database, or it can be directly queried or updated through the database, which is not specifically limited here.
[0068] In this step, the specific implementation process of determining the reference image corresponding to the input image includes: first, preprocessing the input image, and determining the maximum level of tiling according to the resolution of the input image after reprojecting it to the Web map tile. Since there is inconsistency between the input image projection and the Web map tile, for example, the secondary standard data product of satellite remote sensing images is usually UTM projection, and the Web map tile mostly uses WGS84 longitude and latitude, the maximum level is determined according to the resolution of the input image after reprojecting it to the Web map tile, and the input image resolution is required to be not less than the determined maximum level. At the same time, the update range of the Web map tile is determined according to the geographic coordinate range, which is the geographic coordinate range determined by all tile images that have an intersection with the geographic range of the input image after the input image is reprojected to the Web map tile projection under the maximum level of the Web map tile. Then, download the maximum level tile image within the update range. This process can request tile images through the WMTS slice map Web service, or directly query and download tile images from the Web map tile database through programming, and the downloaded tile images can be saved in memory or local hard disk. Finally, the downloaded tile images are spliced and mosaicked into a complete remote sensing image containing geographic coordinates and map projection according to the tile division rules of the Web map tiles and the row and column numbers determined by the geographic coordinate positions, which is used as a reference image.
[0069] It should be noted that most multispectral remote sensing images are currently 4-band unsigned 16-bit, while the tile images stored in the Web map tiles are 3-band RGB color 8-bit. Therefore, in the input image preprocessing process, it is necessary to simultaneously perform RGB band selection and 16-bit to 8-bit processing during the reprojection process. Among them, the 16-bit to 8-bit operation adopts a 2% linear stretching method. For each band, the histogram of the 16-bit image band is first counted, and then the pixel values corresponding to the 2% positions before and after the histogram are found, and then the values between the 2% pixel values before and after are linearly mapped to 0~255 to complete the 8-bit conversion process. The processed input image is an 8-bit, 3-band, band order RGB, map projection consistent with the Web map tile, and contains geographic coordinates. The remote sensing image is stored in a temporary folder on the local hard disk.
[0070] like Figure 5 As shown, from left to right are the input image, the input image reprojected according to the Web map tile, and the reference image mosaicked by the downloaded tiles. It can be seen that the reference image is outsourced from the input image. The reference image is a complete remote sensing image mosaicked by the downloaded tiles from the Web map tile, and the geographic coverage is outsourced from the input image, with 8-bit storage geotif format, 3 bands, band order RGB, including map projection and geographic coordinates, and stored in a temporary folder on the local hard disk.
[0071] Step 103: Perform geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy.
[0072] In an embodiment of the present invention, feature points in the input image and the reference image can be extracted respectively, and feature matching is performed on the feature points of the input image and the feature points in the reference image to obtain corresponding control point data; a corresponding geometric correction model is determined based on the control point data, and the input image is input into the geometric correction model for geometric precision correction to obtain a standard input image that meets the preset geometric alignment accuracy.
[0073] In order to eliminate the geometric position offset that occurs after the web map tiles are updated, resulting in misplaced "patches" and "seams", the input image needs to be geometrically corrected. The specific processing process includes two steps: control point matching and geometric correction, so that the relative geometric accuracy of the output result image (i.e., the standard input image) and the reference image is higher.
[0074] In the process of control point matching, feature points must first be extracted from the input image and the reference image respectively. SIFT (Scale-invariant feature transform) feature extraction is often used for high-resolution remote sensing images. The SIFT algorithm is to find extreme points in the spatial scale and extract their position, scale, and rotation invariants. SIFT features refer to local extreme values that are significantly different from the surrounding and adjacent scale points in the same region in the image scale space, and are the maximum / minimum values of the differential Gaussian function in the local image. SIFT feature points appear as points with grayscale values that are larger or smaller than those of the surrounding points and adjacent scale points in the same region. Then, feature matching is performed on the extracted feature points. Feature matching is the process of finding the same feature points from the SIFT feature points extracted from two images. For each feature point, a coarse matching point is generated for each candidate matching point based on the geographic coordinate position relationship between the images to form a coarse matching point pair. Then, with each pair of rough matching points as the center, the template window (M×M) and the search window (N×N, N>M) are extracted from the two matching images respectively. Here, M is 128 and N is 512. The template window is moved in the search window, and the similarity between each position and the search window is calculated. If the maximum similarity is greater than a certain threshold T, where T is 0.65, the corresponding matching position is taken as the exact matching result of the pair of control points, and the maximum similarity is recorded.
[0075] In the geometric correction step, the geometric correction model is first established using the matched control point data. If the control point data is dense and evenly distributed, the geometric correction uses a triangulated mesh model, which uses all control points to establish a Delaunay triangulation. The error of the triangulated mesh model at the vertices (control points) of the triangulation is 0, and linear interpolation is performed inside each triangle of the triangulated mesh model. Therefore, the denser the control points, the smaller the triangle area and the higher the overall interpolation accuracy. If the number of control points is not dense enough, a second-order polynomial model is used. The second-order polynomial model uses the least squares method to fit a second-order polynomial to all control points, and the obtained model has the smallest average error among all control points.
[0076] The geometric registration accuracy of the input image and the reference image is improved to the sub-pixel level by using a part of the control points obtained in the matching step as check points to determine whether the average error of the calculation result is less than 1. After the geometric precision correction model is established, before the geometric precision correction of the input image is performed, the control point precision analysis is added. If the original precision of the input image is less than 1 (sub-pixel level), the geometric precision correction is not performed.
[0077] Step 104: Perform cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determine a clear surface area of the standard input image based on the mask image.
[0078] In an embodiment of the present invention, the standard input image and the reference image may be input into a cloud and cloud shadow detection model for change detection, and a guided filtering model may be used to correct the boundary between clouds and shadows to obtain a mask image of the standard input image.
[0079] In this step, the purpose of cloud and cloud shadow detection is to exclude the cloud and cloud shadow areas in the Wed map tile update data, and obtain a mask image marked with clouds and cloud shadows. If the satellite remote sensing image contains cloud and cloud shadow detection results, the mask image can be directly generated. Otherwise, it is necessary to determine the cloud and cloud shadow detection model (i.e., cloud and cloud shadow detection algorithm) for the standard input image.
[0080] In the cloud and cloud shadow detection model, it is assumed that the reference image is clear and cloudless. Change detection is performed on the input image and the reference image, and the uniformly bright and dark areas in the changed parts are extracted to identify them as clouds and shadows under the clouds. The IR-MAD (re-weighted Multivariate Alteration Detection transformation) transformation based on multi-source canonical correlation analysis is used to automatically extract the multi-source change analysis variables (MAD variables) of the two images. The use of IR-MAD transformation further improves the accuracy and stability of the MAD algorithm. For the MAD variables between the input image and the reference image, the areas of change such as clouds and shadows under the clouds correspond to larger pixel values. All the change areas can be extracted by presetting the threshold. Assuming that the RGB colors of clouds and shadows under the clouds are monochrome white and black, most of the falsely detected change surfaces can be excluded according to the color threshold. Finally, for the extracted cloud and shadow areas under the clouds, the guided filtering model (i.e., guided filtering algorithm) is used to correct the boundaries to obtain a mask image that is more consistent with the clouds and shadows under the clouds in the standard input image.
[0081] In the guided filtering algorithm, the standard input image is used as the guided image L(x, y), the mask image is used as the filtered image V(x, y), and the output image is recorded as The filtering result at the pixel (x, y) is expressed as a weighted average:
[0082]
[0083] Assume that the guided filter is between the guided image L(x,y) and the filtered output In between is a local linear model:
[0084]
[0085] By minimizing the following window k The cost function is:
[0086]
[0087] Get the local linear coefficient a k ,b k The output image of the guided filter The corrected cloud boundaries can be extracted using a simple threshold.
[0088] The mask image has the same pixel size as the standard input image and is in the format of a single-band 8-bit image, where pixel values of 0 correspond to clear surfaces and non-zero values correspond to excluded areas, including clouds and cloud shadows, and background areas.
[0089] Step 105: using the reference image as a base image, determining a target stitching line between the input image and the reference image; replacing the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain a target input image.
[0090] In an embodiment of the present invention, the reference image may be used as a base map of the standard input image, a multi-band array composed of bands between the input image and the reference image is determined based on a preset remote sensing image stitching line search model, and a target stitching line is determined based on the multi-band array; the implementation process of replacing the exclusion area in the input image with the corresponding area of the reference image based on the target stitching line includes: determining the target stitching line as a boundary between a clear surface area and an exclusion area in the mask image; and replacing the exclusion area in the input image with the corresponding area of the reference image based on the boundary to obtain a target input image.
[0091] Specifically, the purpose of stitching line search is to replace the mask boundary with the natural boundary of the object coverage, so as to eliminate the stitching seam phenomenon after the web map tile is updated. In the process of stitching line search, the remote sensing image stitching line search model (remote sensing image stitching line search algorithm) is used to find the target stitching line between the standard input image and the reference image. Figure 6As shown, the maximum search radius is defined as r (5 < r < 20) pixels. The search area is the area within r pixels from the clear surface boundary corresponding to the mask image in the standard input image. The search direction v is parallel to the area boundary, and the input data for the search is the multi-band array composed of the standard input image and the reference image bands within the search area. Further, the obtained target stitching line is used to replace the boundary between the clear surface and the excluded area in the mask image. Its morphological effect is equivalent to performing an erosion operation with a maximum radius of r on the clear surface area in the mask image. The pixels in the standard input image marked as the excluded area in the mask image are filled with the pixels of the reference image. After processing, the pixel positions in the standard input image corresponding to the excluded area marked in the mask image are filled with the pixels of the reference image at the corresponding geographical coordinates. At this time, it is stated that the excluded area in the input image is replaced by the corresponding area of the reference image, and the geometric offset and stitching line problems between the clear surface area and the corresponding area of the reference image are eliminated, and only the differences in color and brightness remain.
[0092] Therefore, after step 105 is executed, it further includes: inputting the target input image and the reference image into a preset remote sensing image equalization and color correction model based on Poisson fusion for equalization and color correction processing to obtain the processed target input image. Among them, the area corresponding to Poisson fusion in the remote sensing image equalization and color correction model is the clear surface area of the mask image of the standard input image. Using the pixels of the reference image corresponding to the outside of the boundary of the clear surface area as the guiding vector, the guiding vector field is calculated towards the inside of the clear surface area to achieve equalization and color correction processing.
[0093] Specifically, based on the remote sensing image equalization and color correction model (remote sensing image equalization and color correction algorithm) based on Poisson fusion, the target input image and the mask image are used as inputs. The area of Poisson fusion is the clear surface area corresponding to the mask image. Using the pixels of the reference image outside the boundary of the clear surface area as the guiding vector, the guiding vector field is calculated towards the inside of the clear area. The pixel values of the target input image corresponding to the reference image area remain unchanged after Poisson fusion processing, and the differences in color and brightness between the clear surface area and the area of the reference image are eliminated.
[0094] As Figure 7 is a schematic diagram of the principle of the remote sensing image equalization and color correction algorithm using the Poisson fusion method. Denote S as the pixel area of the reference image filled in the target input image. This area does not participate in the Poisson fusion process and the pixel values remain unchanged. Denote Ω as the area of the target input image to be equalized and color corrected. The boundary pixels of the area are denoted Denote f * as the scalar function of the range of S minus Ω. Denote f as the unknown scalar function of the range of Ω. The guiding vector field calculated from g is denoted as V. The purpose of Poisson fusion is to solve for f so that the patched area has the best consistent transition with S, and operationally, the internal gradient change of Ω is minimized. Denote is the gradient operation, then we have And at the border there is remember is the Laplace operator, and it reaches its minimum value when the second-order partial derivative of the gradient is 0. According to the Lagrange equation, Δf| Ω = 0, which is a guide vector domain minimization problem. Its solution is the unique solution of the Poisson equation under Dirichlet boundary conditions, namely, Δf| Ω =divV, and also at the boundary in is the divergence of V = (u, v). The purpose of Poisson fusion is to solve the correction function within Ω. Make Therefore, the correction required within Ω is Through the border With error f * -g is used as the boundary condition and the solution is obtained by interpolation into Ω. And at the border there is
[0095] For the processed target input image, further determine whether the processing result meets the update standard of the Web map tile, that is, there is no obvious "patch", "stitching line" and other geometric position offset and color brightness inconsistency, and the clear surface part of the target input image is completely integrated with the filled reference image area, then subsequent tiling processing can be performed.
[0096] Step 106: tile the target input image, and use the generated tile data to update the original map tile data in the network map reference base map database.
[0097] In an embodiment of the present invention, a target input image can be processed into tile image data of a preset format according to a preset tile division rule; wherein the tile image of the largest level in the tile image data is obtained by directly segmenting the target input image, the tile image with incomplete coverage in the tile image data is obtained by filling the reference image, and the upper layer tiles in the tile image data are obtained by merging and resampling the lower layer tiles; and the tile image data is used to update the original map tile data in the network map reference base map database.
[0098] Specifically, the tiling process includes: tiling the processed input image according to the tile division rule of the Web map tile, processing the target input image into a tile image of a fixed size (such as 256*256 pixel size) in jpg or png format, and using the generated tile image data to update the Web map reference base map database to complete the update of the Web map tile data.
[0099] The Web map tile update processing method provided by the embodiment of the present invention performs geometric precision correction on the input image through a reference image to obtain a standard input image that meets the preset geometric registration accuracy, and performs cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determines the clear surface area of the standard input image based on the mask image, and uses the reference image as the base map to determine the target stitching line between the input image and the reference image, and replaces the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain the target input image; the target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database. This method can make the generated map tile data have higher quality Web map tile update processing.
[0100] Corresponding to the Web map tile update processing method provided above, the present invention also provides a Web map tile update processing device. Since the embodiment of the device is similar to the above method embodiment, the description is relatively simple. For the relevant parts, please refer to the description of the above method embodiment. The embodiment of the Web map tile update processing device described below is only illustrative.
[0101] Please refer to Figure 8 As shown, it is a structural diagram of a Web map tile update processing device provided by an embodiment of the present invention.
[0102] The Web map tile update processing device of the present invention comprises the following parts:
[0103] The remote sensing image acquisition unit 801 is used to acquire an input image; the input image is a remote sensing image to be tiled;
[0104] The reference image determination unit 802 is used to determine a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data;
[0105] A geometric precision correction unit 803 is used to perform geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy;
[0106] a cloud and cloud shadow detection unit 804, configured to perform cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determine a clear surface area of the standard input image based on the mask image;
[0107] The image stitching line determining unit 805 is used to determine a target stitching line between the input image and the reference image using the reference image as a base image; based on the target stitching line, the excluded area in the input image is replaced with the corresponding area of the reference image to obtain a target input image;
[0108] The tile processing unit 806 is used to perform tile processing on the target input image and use the generated tile data to update the original map tile data in the network map reference base map database.
[0109] The Web map tile update processing device provided in the embodiment of the present invention performs geometric precision correction on the input image through a reference image to obtain a standard input image that meets the preset geometric registration accuracy, and performs cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, determines the clear surface area of the standard input image based on the mask image, takes the reference image as the base map, determines the target stitching line between the input image and the reference image, replaces the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain the target input image; performs tile processing on the target input image, and uses the generated tile data to update the original map tile data in the web map reference base map database. This method can make the generated map tile data of higher quality.
[0110] Corresponding to the Web map tile update processing method provided above, the present invention also provides an electronic device. Since the embodiment of the electronic device is similar to the above method embodiment, the description is relatively simple. For the relevant parts, please refer to the description of the above method embodiment. The electronic device described below is only exemplary. Fig. 9As shown, it is a schematic diagram of the physical structure of an electronic device disclosed in an embodiment of the present invention. The electronic device may include: a processor (processor) 901, a memory (memory) 902 and a communication bus 903, wherein the processor 901 and the memory 902 communicate with each other through the communication bus 903 and communicate with the outside through the communication interface 904. The processor 901 can call the logic instructions in the memory 902 to execute the Web map tile update processing method. The method comprises: obtaining an input image; the input image is a remote sensing image to be tiled; determining a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing original map tile data; performing geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy; performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determining a clear surface area of the standard input image based on the mask image; using the reference image as a base map, determining a target stitching line between the input image and the reference image; replacing an excluded area in the input image with a corresponding area of the reference image based on the target stitching line to obtain a target input image; performing tile processing on the target input image, and using the generated tile data to update the original map tile data in a network map reference base map database.
[0111] In addition, the logic instructions in the above-mentioned memory 902 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 can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several 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 method described in each embodiment of the present invention. The aforementioned storage medium includes: a storage chip, a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and other media that can store program codes.
[0112] On the other hand, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a processor-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the Web map tile update processing method provided by the above-mentioned method embodiments. The method comprises: obtaining an input image; the input image is a remote sensing image to be tiled; determining a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing original map tile data; performing geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy; performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determining a clear surface area of the standard input image based on the mask image; using the reference image as a base map, determining a target stitching line between the input image and the reference image; replacing an excluded area in the input image with a corresponding area of the reference image based on the target stitching line to obtain a target input image; performing tile processing on the target input image, and using the generated tile data to update the original map tile data in a network map reference base map database.
[0113] On the other hand, an embodiment of the present invention further provides a processor-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it is implemented to execute the Web map tile update processing method provided by the above embodiments. The method includes: obtaining an input image; the input image is a remote sensing image to be tiled; determining a reference image corresponding to the input image; the reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data; performing geometric precision correction on the input image based on the reference image to obtain a standard input image that meets the preset geometric registration accuracy; performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determining the clear surface area of the standard input image based on the mask image; using the reference image as the base map, determining the target stitching line between the input image and the reference image; replacing the exclusion area in the input image with the corresponding area of the reference image based on the target stitching line to obtain the target input image; performing tile processing on the target input image, and using the generated tile data to update the original map tile data in the web map reference base map database.
[0114] The processor-readable storage medium can be any available medium or data storage device that can be accessed by the processor, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NANDFLASH), solid-state drive (SSD)), etc.
[0115] 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 displayed as units may or may not be physical units, that is, 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. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0116] 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, it 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 described in each embodiment or some parts of the embodiments.
[0117] 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 method for updating a web map tile, characterized in that: include: Acquire an input image; the input image is a remote sensing image to be tiled; Determining a reference image corresponding to the input image; The reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data; Performing geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy; Performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determining a clear surface area of the standard input image based on the mask image; Taking the reference image as a base image, determining a target stitching line between the input image and the reference image; replacing the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain a target input image; The target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database.
2. The Web map tile update processing method according to claim 1, characterized in that: Before tiling the target input image, the method further includes: The target input image and the reference image are input into a preset remote sensing image even-lighting and color-matching model based on Poisson fusion for even-lighting and color-matching processing to obtain the processed target input image; wherein the corresponding Poisson fusion area in the remote sensing image even-lighting and color-matching model is the clear surface area of the mask image of the standard input image, and the pixels of the reference image corresponding to the boundary of the clear surface area are used as guide vectors, and the guide vector domain is calculated toward the inside of the clear surface area to achieve even-lighting and color-matching processing.
3. The Web map tile update processing method according to claim 1, characterized in that: Determining a reference image corresponding to the input image specifically includes: Determining a maximum level of tiling according to a resolution of the input image after reprojection onto network map tiles; the resolution of the input image is not less than the determined maximum level; Determining corresponding geographic coordinate ranges based on all network map tile images in the reference image that intersect with the geographic range of the input image; Determine the update range of the network map tiles according to the geographic coordinate range, obtain the tile image of the maximum level within the update range, splice the tile image of the maximum level into a complete remote sensing image; and use the complete remote sensing image as a reference image corresponding to the input image.
4. The Web map tile update processing method according to claim 1, characterized in that: The input image is geometrically precisely corrected based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy, specifically comprising: Extracting feature points from the input image and the reference image respectively, and performing feature matching on the feature points of the input image and the feature points in the reference image to obtain corresponding control point data; A corresponding geometric correction model is determined based on the control point data, and the input image is input into the geometric correction model for geometric precision correction to obtain a standard input image that meets a preset geometric registration accuracy.
5. The Web map tile update processing method according to claim 1, characterized in that: Performing cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image specifically includes: inputting the standard input image and the reference image into a cloud and cloud shadow detection model to perform change detection, and using a guided filtering model to correct the boundary between clouds and shadows to obtain the mask image of the standard input image.
6. The Web map tile update processing method according to claim 1, characterized in that: Taking the reference image as a base image, determining a target stitching line between the input image and the reference image specifically includes: Using the reference image as the base image of the standard input image, determining a multi-band array composed of bands between the input image and the reference image based on a preset remote sensing image stitching line search model, and determining a target stitching line based on the multi-band array; Replacing the excluded area in the input image with the corresponding area of the reference image based on the target stitching line specifically includes: The target stitching line is determined as a boundary between a clear surface area and an exclusion area in the mask image; and based on the boundary, the exclusion area in the input image is replaced with a corresponding area of the reference image to obtain a target input image.
7. The Web map tile update processing method according to claim 1, characterized in that: The target input image is tiled, and the generated tile data is used to update the original map tile data in the network map reference base map database, specifically including: The target input image is processed into tile image data in a preset format according to a preset tile division rule, and the tile image data is used to update the original map tile data in the network map reference base map database.
8. A Web map tile update processing device, characterized in that: include: A remote sensing image acquisition unit, used for acquiring an input image; The input image is a remote sensing image to be tiled; A reference image determining unit, used to determine a reference image corresponding to the input image; The reference image is a remote sensing image containing geographic coordinates and map projection obtained by preprocessing the original map tile data; A geometric precision correction unit, configured to perform geometric precision correction on the input image based on the reference image to obtain a standard input image that meets a preset geometric registration accuracy; a cloud and cloud shadow detection unit, configured to perform cloud and cloud shadow detection on the standard input image to obtain a mask image of the standard input image, and determine a clear surface area of the standard input image based on the mask image; An image stitching line determining unit, configured to determine a target stitching line between the input image and the reference image using the reference image as a base image; replacing the excluded area in the input image with the corresponding area of the reference image based on the target stitching line to obtain a target input image; The tile processing unit is used to perform tile processing on the target input image and use the generated tile data to update the original map tile data in the network map reference base map database.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the Web map tile update processing method as described in any one of claims 1 to 7 are implemented.
10. A processor-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the Web map tile update processing method as claimed in any one of claims 1 to 7 are implemented.