High-performance visualization method, system and electronic equipment for large-scale remote sensing image data

Abstract

The present invention discloses a high-performance visualization method, system and electronic device for large-scale remote sensing image data, the method comprising: S1, constructing a remote sensing image sky map correspondence library; S2, collecting remote sensing image data, extracting remote sensing image resolution corresponding to the sky map level and the corresponding optimal GeoSOT level, slicing the remote sensing image data according to the grid of the optimal GeoSOT level and storing it as optimal level slice data; S3, performing grid slicing and downsampling processing step by step according to the GeoSOT level starting from the optimal GeoSOT level, and obtaining slice data of each GeoSOT level in turn; S4, filtering all slice data of the corresponding GeoSOT level based on the sky map level selected by the query viewport and rendering to generate a visualization image. The present invention constructs a remote sensing image sky map correspondence library and uses it to perform optimal GeoSOT level adaptation of remote sensing image data, and performs GeoSOT level slice expansion to adapt to more query viewport resolutions, greatly improving visualization quality and efficiency.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing data processing, and in particular to a high-performance visualization method, system and electronic equipment for large-scale remote sensing image data. Background Art

[0002] Remote sensing images are images of the Earth's surface acquired through remote sensing technology, containing spectral information about surface features. With the development of cloud computing, remote sensing, and sensor technologies, remote sensing science and technology are becoming increasingly important and are widely used in various fields, such as air pollution, environmental protection, ecological restoration, and urban planning. Remote sensing image data visualization technology plays a crucial role in this field. Tianditu, as a national geographic information public service platform, integrates resources from geographic information departments at all levels to provide authoritative and unified online services. It is often used as the base map for remote sensing image visualization services.

[0003] Remote sensing image data visualization typically adheres to OGC standards, such as WMS and WMTS. WMS renders remote sensing images into images and returns them via HTTP services, while WMTS slices images into tiles and pre-renders them for storage, improving access efficiency. In addition, non-OGC visualization methods, such as TMS, share similar principles but are less standardized. These visualization methods suffer from three issues: First, access efficiency is relatively low. This can lead to significant performance bottlenecks when the data volume is large, such as with high-resolution images or low network bandwidth. Second, the storage method is monotonous and difficult to scale. Both the image itself and the resulting image slices must be stored on disk. File organization and management rely on the operating system's file system, which cannot effectively utilize distributed file systems. Read and write performance depends on host performance, and there is a risk of single points of failure. Third, resource usage is significant, with memory usage increasing with the addition of services. With the advancement of remote sensing technologies, the resolution and size of images available are increasing. These three issues significantly restrict the use of imagery, necessitating an efficient visualization solution that can be deployed in a distributed manner and consumes minimal resources. Summary of the Invention

[0004] The purpose of the present invention is to solve the technical problems of low efficiency, single storage and difficulty in expansion, and serious resource occupation in existing remote sensing image data visualization services. It provides a high-performance visualization method, system and electronic equipment for large-scale remote sensing image data. Through GeoSOT grid coding and calculation, distributed file system, and map color rendering, it can achieve fast access, flexible storage and low resource consumption. Regardless of the resolution or size of the remote sensing image, it can achieve reliable storage, fast retrieval and efficient rendering visualization.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A high-performance visualization method for large-scale remote sensing image data, comprising:

[0007] S1. Construct a remote sensing imagery sky map correspondence library including the sky map level, GeoSOT level, and the level correspondence; the sky map level includes the sky map resolution, and the GeoSOT level includes the GeoSOT size;

[0008] S2. Collect remote sensing image data, extract remote sensing image resolution corresponding to the adapted sky and earth map level and the corresponding optimal GeoSOT level, slice the remote sensing image data according to the grid of the optimal GeoSOT level and store it as optimal level slice data;

[0009] S3, starting from the optimal GeoSOT level, grid slicing and downsampling are performed step by step according to the GeoSOT level to obtain slice data of each GeoSOT level in turn;

[0010] S4. Based on the selected sky and earth map level of the query viewport, all slice data corresponding to the GeoSOT level are filtered and rendered to generate a visualization image.

[0011] In order to better implement the present invention, step S3 also includes the following method:

[0012] The optimal level slice data and the slice data of each GeoSOT level are GeoSOT grid coded according to the grid.

[0013] Preferably, in step S1, the sky map resolution of the sky map level increases sequentially from level to level, and the sky map resolution of the GeoSOT level decreases sequentially from level to level, and the hierarchical correspondence is a one-to-one correspondence between the sky map level and the GeoSOT level; the highest sky map resolution of the sky map level corresponds to the lowest GeoSOT size of the GeoSOT level, and the lowest sky map resolution of the sky map level corresponds to the highest GeoSOT size of the GeoSOT level.

[0014] Preferably, in step S2, the sky map resolution of the sky map level is adapted according to the extracted remote sensing image resolution and the sky map level is adapted, and the corresponding GeoSOT level is found as the optimal GeoSOT level based on the remote sensing image sky map correspondence library; mask extraction slicing and grid coding processing are performed based on the grid of the optimal GeoSOT level to obtain the optimal level slice data containing grid coding information.

[0015] Preferably, in step S3, the GeoSOT size is increased by one level compared to the optimal GeoSOT level based on the GeoSOT level, and mask extraction slicing, grid coding, and downsampling processing are performed based on the grid of the increased GeoSOT level to obtain slice data containing grid coding information and the increased GeoSOT level; the GeoSOT level is increased step by step to obtain slice data of each GeoSOT level.

[0016] Preferably, step S3 further includes the following method:

[0017] Based on the GeoSOT level, the GeoSOT size is reduced by one level compared to the optimal GeoSOT level. Mask extraction, slicing, grid encoding, and upsampling are performed on the grid of the reduced GeoSOT level to obtain slice data containing grid coding information and reduced GeoSOT level by one level; the GeoSOT level is gradually reduced to obtain slice data of the remaining GeoSOT levels.

[0018] Preferably, a distributed file system is used to store all slice data in a distributed manner.

[0019] Preferably, in step S4, all slice data at the GeoSOT level are screened out and the data content is extracted for standardization and color rendering to generate a rendered visual image.

[0020] A high-performance visualization system for large-scale remote sensing image data includes a data acquisition module, a slice code storage system, and a query viewport system. The slice code storage system has a remote sensing image sky map correspondence library, which includes a sky map level, a GeoSOT level, and a level correspondence. The sky map level includes the sky map resolution, and the GeoSOT level includes the GeoSOT size. The data acquisition module is used to collect remote sensing image data and input it into the slice code storage system. The slice code storage system extracts the remote sensing image data and the remote sensing image resolution corresponds to the sky map level. The remote sensing image data is sliced and processed according to the grid of the optimal GeoSOT level and stored as optimal level slice data. The slice encoding storage system performs grid slicing and downsampling processing step by step according to the GeoSOT level, starting from the optimal GeoSOT level, to obtain slice data for each GeoSOT level in sequence. The query viewport system includes a filtering module and a rendering module. The filtering module is used to select the sky and earth map level and filter all slice data of the corresponding GeoSOT level and input them into the rendering module. The rendering module uses all slice data to render and generate a visual image. This steadily improves query efficiency, while storage can be easily expanded. In addition, it does not need to be preloaded into memory, reducing resource usage.

[0021] An electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one processor to perform the steps of the high-performance visualization method for large-scale remote sensing image data of the present invention.

[0022] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0023] (1) The present invention extracts the remote sensing image resolution corresponding to the adaptive sky and earth map level and the corresponding optimal GeoSOT level from the collected remote sensing image data, then performs grid slicing and downsampling processing step by step according to the GeoSOT level and stores all the slice data respectively, and filters all the slice data corresponding to the sky and earth map level selected by the query viewport and renders to generate a visual image; by dividing the remote sensing image level into multiple grids, the distributed storage of the image itself is realized, and the rapid query of the image range is realized through grid calculation; the queried grid is used to generate a picture in real time through map color matching, and finally the picture is rendered on the map, so as to quickly realize data query and display without occupying additional resources.

[0024] (2) The present invention provides a high-performance visualization method, system, and electronic device for large-scale remote sensing image data. Through GeoSOT grid coding and calculation, distributed file system, and map color rendering, it can achieve fast access, flexible storage, and low resource consumption. Regardless of the resolution or size of the remote sensing image, reliable storage, fast retrieval, and efficient rendering visualization can be achieved.

[0025] (3) The present invention constructs a remote sensing image sky map correspondence library and uses it to perform optimal GeoSOT level adaptation of remote sensing image data, realizing the optimal visualization adaptation strategy of query viewport resolution and remote sensing image resolution, and then expands GeoSOT level slices to adapt to more query viewport resolutions, thereby improving the remote sensing image data screening, reading and downloading speed and subsequent rendering efficiency, greatly improving the visualization quality and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A flowchart of the method for high-performance visualization of large-scale remote sensing image data according to the present invention;

[0027] Figure 2 Schematic diagram of the principle of grid division and encoding in the embodiment;

[0028] Figure 3 Schematic diagram of the principle of pyramid-like slicing starting from the optimal GeoSOT level in the embodiment;

[0029] Figure 4 Schematic diagram of the principle of query viewport filtering slice data and rendering under query request in an embodiment;

[0030] Figure 5 This is a principle structural block diagram of the high-performance visualization system for large-scale remote sensing image data of the present invention. DETAILED DESCRIPTION

[0031] Below in conjunction with embodiment, the present invention is described in further detail:

[0032] Example

[0033] like Figure 1 As shown, a high-performance visualization method for large-scale remote sensing image data includes:

[0034] S1. Construct a remote sensing imagery sky map correspondence library including the sky map level, GeoSOT level, and the level correspondence; the sky map level includes the sky map resolution, and the GeoSOT level includes the GeoSOT size.

[0035] In step S1, the sky map resolution of the sky map level increases in sequence from level to level, and that of the GeoSOT level decreases in sequence from level to level, and the hierarchical correspondence is a one-to-one correspondence between the sky map level and the GeoSOT level; the highest sky map resolution of the sky map level corresponds to the lowest GeoSOT size of the GeoSOT level, and the lowest sky map resolution of the sky map level corresponds to the highest GeoSOT size of the GeoSOT level.

[0036] Assume the map scale is 1:100 (to simplify calculations, assume a pixel is 1 cm long and 1 cm wide), and the query viewport pixel resolution is 100*100 (100 pixels long and 100 meters wide). This means that one pixel represents a geographic range of 100 meters long and 100 meters wide. The viewport should display a geographic range of 100 pixels long and 100 meters wide = 10,000 meters. Assume there are four map images with different resolutions (1 meter, 10 meters, 100 meters, and 1000 meters). This means that one pixel represents 1 meter, 10 meters, 100 meters, and 1000 meters, respectively. There are four ways to load imagery: The first way is to get a resolution of 1 meter. You need to load 10,000 pixels with a length and width of 10,000 meters / 1 meter = 10,000 pixels. According to the three bands of RGB, each pixel in each band can be expressed by one Byte (0-255). The size of the converted image is: 10,000*10,000*3 / 1024 / 1024 = 286MB. The second way is to get a resolution of 10 meters. You need to load 10,000 meters with a length and width of 10,000 meters / 10 =1000 pixels, and the image size converted to 1000*1000*3 / 1024 / 1024=2.86MB according to the same method; the third method, to obtain a resolution of 100 meters, needs to load 10000 meters / 100 meters = 100 pixels in length and width, and the image size converted to 100*100*3 / 1024=29KB according to the same method; the fourth method, to obtain a resolution of 1000 meters, needs to load 10*10*3=300B in length and width. In other words, no matter what the resolution, the front-end map (the query viewport of the present invention selects the world map) can be loaded, but it can be seen from the size of the image that the lower the image resolution, the smaller the actual generated image. For the front-end loading images through the network, the smaller the image, the faster it loads in theory.

[0037] Let's analyze the above four methods. The first method is to display 10,000 pixels in length and width in an area of 100 pixels in length and width, which is equivalent to displaying 100 pixels per pixel, like compressing a picture with a resolution of 10,000*10,000 to 100*100, which is very fine; the second method is equivalent to displaying 10 pixels per pixel, which will be finer; the third method does not require any compression or stretching, and highly restores the true situation of the image; the fourth method is equivalent to displaying 0.1 pixels per pixel, like stretching a 100*100 pixel picture to 1000*1000, which will show obvious graininess and the display will be very rough.

[0038] The above analysis shows that, given a fixed viewport size and map scale, lower image resolution is best for faster loading, while higher resolution is best for better display quality. This leads to the conclusion that an image resolution exactly equal to the viewport pixel resolution is optimal. When it comes to visualization, display quality is paramount. A slightly larger image size is acceptable, as long as the display remains distortion-free and even slightly refined.

[0039] Let's examine the impact of map scale on loading efficiency. Let's assume the viewport size is 100 pixels long and wide, and the map scale is 1:100. Assuming the viewport size remains unchanged and the map scale is changed to 1:10, the map will display a geographic area of 100 meters long and 10 meters wide. The optimal approach is to load a 10-meter resolution image, which converts to an image size of 1000 / 10*1000 / 10*3 / 1024 = 29KB. Similarly, if the scale is changed to 1:1000, the same result is achieved. The same applies to other scales. Therefore, we can conclude that, given a fixed viewport, map scale does not affect the size of the loaded image.

[0040] Next, let's analyze the impact of viewport size on loading efficiency. The viewport size is set to 100 pixels in both length and width. This isn't always possible in practice. Currently, the mainstream screen resolution is 2K (2560*1440), so we can set the maximum viewport to 2K. If we can achieve smooth loading with a 2K viewport, then other smaller viewports will have no problems. Let's analyze the scenario above, changing the viewport size to 2K and loading using the third method. The range that the viewport needs to display is: 2560*100=256000 meters in length and 1440*100=144000 meters in width. The size that needs to be converted into an image is: 256000 / 100*144000 / 100*3 / 1024 / 1024=10.5MB, which is larger than the original. In fact, the image size increases by the same amount as the number of viewport pixels increases: 2560*1440 / (100*100)=10.5MB / 29KB.

[0041] That is to say, on a 2K screen, the maximum loaded image size is 10.5MB. If the bandwidth is 100Mbps, then the time required to download this image is about 10.5M / (100Mbps / 8) = 0.84 seconds. Because the image is relatively large, the post-rendering (using H5 rendering) is also relatively time-consuming, about 0.1 seconds. Without considering other bandwidth occupants, the total display time is less than 1 second. The resulting effect is that the interface freezes for about 1 second, and then suddenly displays. The intuitive feeling is not only slow, but also not smooth. However, corresponding optimizations can be made: because different browsers have a concurrency of 4 to 6 for each domain name when loading images, a large image can be divided into multiple small images, each of which is loaded concurrently. The visual effect is that the small images are presented one after another during a waiting time of about 1 second, and the effect is relatively smooth. Consider dividing the large image into small square images of the same length and width:

[0042] (1) Divide the width of the image into two parts, so the width and height of the small image are 1440 / 2 = 720 pixels, and the length is 2560 / 720 = rounded up to 4. A total of 2*4 = 8 images are needed, and the size of each image is 10.5MB / 8 = 1.3MB;

[0043] (2) Divide the width of 1440 into 3 parts. The width and height of the small picture are 1440 / 3 = 480 pixels, and the length is 2560 / 480 = rounded up to 6. A total of 3*6 = 18 pictures are needed, and the size of each picture is 10.5MB / 18 = 583KB.

[0044] (3) Divide the width of 1440 into 4 parts, then the length and width of the small picture are 1440 / 4 = 360 pixels, and the length is 2560 / 360 = rounded up to 8, so a total of 4*8=32 pictures are needed, and the size of each picture is 10.5MB / 32=328KB;

[0045] (4) Divide the width of 1440 into 5 parts, then the length and width of the small picture are 1440 / 5 = 288 pixels, and the length is 2560 / 288 = rounded up to 9. A total of 5*9 = 45 pictures are needed, and the size of each picture is 10.5MB / 45 = 233KB

[0046] …

[0047] Considering that the size of small images cannot be too large, (2), (3), and (4) meet the requirements. Considering that the number of concurrent browsers is limited (only 4-6), each concurrent loading of images cannot be too many, otherwise the sum of multiple small images will be too large and loading will be very slow, so (2) and (3) are more in line with the requirements. Consider another situation: the resolution of images varies, and it is generally impossible to be the same as the pixel resolution corresponding to the map scale. In order to meet the display effect, images with a resolution greater than or equal to the pixel resolution will be loaded. If the image resolution is 7 meters and 12 meters, and the pixel resolution is 10 meters, then a 7-meter image will be loaded. According to the above calculation method, the generated image size is: 100*10 / 7*100*10 / 7*3 / 1024=60KB. If we look at a similar situation on a 2K screen, (2) in this case, the size of a small picture can be roughly estimated to be 583*10 / 7=833KB, and (3) is estimated to be 328*10 / 7=469KB. Obviously, (3) has a larger margin and is a more desirable method.

[0048] Based on the above analysis, there is actually a corresponding relationship between viewport size, map scale, pixel resolution, image resolution, image size, number of split images, and split image size. Some examples are shown in Table 1 below (the viewport size is set to 2K):

[0049]

[0050]

[0051] Table 1

[0052] As you can see, if the scale changes are small, the image size can be kept within a fixed range. In fact, if the scale is sufficiently continuous, such as 1:10, 1:9, 1:8..., the image size can even be kept constant.

[0053] In fact, the map zoom level setting can ensure that the scale interval is within a certain range. Take Tiandi Map as an example. It is divided into 24 levels from 0 to 23, with a scale ranging from 1:279541132 to 1:33. On a 2K screen, the pixel resolution ranges from 78184 meters to 0.009 meters. Some examples are shown in Table 2:

[0054]

[0055]

[0056] Table 2

[0057] Next, let's examine GeoSOT and establish its relationship with small images. GeoSOT is an image segmentation method that can be used to segment large images as described above. The goal is to select a GeoSOT segmentation level for different image resolutions that produces images close to the 328KB size discussed above.

[0058] By mapping GeoSOT level 25 to Tiandi Map level 23, GeoSOT level 24 to Tiandi Map level 22, and GeoSOT level 23 to Tiandi Map level 21, we can obtain the following List 3 (partial example):

[0059]

[0060]

[0061] Table 3

[0062] Considering that the actual image resolution does not strictly correspond to the pixel resolution, the above table can be expressed in a range format (in the following table, (xy] indicates a value range greater than x and less than or equal to y):

[0063]

[0064]

[0065] Table 4

[0066] Then calculate the size range of the generated image based on the range boundary (in the following table, (xy] indicates the value range is greater than x and less than or equal to y):

[0067]

[0068] Table 5

[0069] According to Table 5 (Table 5 is part of the remote sensing image sky map correspondence library), combined with the above discussion, under the 2K viewport, the display effect is best when the map loading and the image resolution are consistent with the pixel resolution.

[0070] S2. Collect remote sensing image data and extract the remote sensing image resolution corresponding to the adapted world map level and the corresponding optimal GeoSOT level (for example, the remote sensing image resolution is 0.01 meters, which falls within the range expressed by the 22nd level of the world map, and the corresponding geosot level is 24). Slice the remote sensing image data according to the grid of the optimal GeoSOT level and store it as optimal level slice data.

[0071] In some embodiments, the resolution of the extracted remote sensing image is adapted to the SkyMap resolution of the SkyMap level, and the SkyMap level is adapted. Based on the remote sensing image SkyMap correspondence database, the corresponding GeoSOT level is found as the optimal GeoSOT level. Mask extraction, slicing, and grid coding are performed based on the grid of the optimal GeoSOT level to obtain optimal level slice data containing grid coding information.

[0072] S3. According to the GeoSOT level, grid slicing and downsampling are performed step by step starting from the optimal GeoSOT level (to facilitate the generation of low-level slice data, until level 0 in Table 5), and slice data of each GeoSOT level is obtained in turn. The downsampling rate of this embodiment is 0.5, and the length and width resolution becomes 1 / 2 of the original. Through such a downsampling operation, not only a new resolution is generated, but also the generated slice size and size are guaranteed to be consistent with the original: for example, the original slice resolution is 100*100, and the merged resolution is 200*200. The resolution remains unchanged, the size becomes twice the original, and the size also becomes twice the original. At this time, a downsampling operation of 0.5 is performed to change the resolution to 100*100, keeping the size and size unchanged. Then, a lower-level slice is generated based on the generated first-level slice, which is consistent with the above method. This cycle is repeated until level 0. The remote sensing image that is loaded by the present invention and consistent with the pixel resolution is optimal (the optimal GeoSOT level has been obtained), such as Figure 3 As shown, this embodiment constructs a pyramid. When the map scale changes, it is necessary to be able to provide remote sensing images of corresponding resolutions. If only one level is divided according to the first step, it can only correspond to one resolution, which cannot meet the demand. First, a lower-level slice is generated: the main process is to merge the original slices and downsample, and then generate a lower-level slice based on the generated first-level slice, and so on. The merging method is roughly to merge the four adjacent slices in the upper, lower, left and right directions into one. According to the rules of Geosot coding, only slices with the same prefix can be merged. Each level of Geosot coding has a fixed number of bits.

[0073] In some embodiments, the present invention further comprises the following method:

[0074] The optimal level slice data and the slice data of each GeoSOT level are processed by GeoSOT grid coding according to the grid. Figure 2 As shown in the following table, GeoSOT encoding rules: only those with the same prefix can be merged. Each level of GeoSOT encoding has a fixed number of bits: the maximum encoding level is 32, divided into four sections, namely 9-bit degree grid code, 6-bit sub-grid code, 6-bit second grid code and 11-bit sub-second grid code. As shown below:

[0075] dddddddd-mmmmmm-ssssss.uuuuuuuuuuu

[0076] 9-digit degree code, 6-digit grade code, 6-digit second code, 11-digit sub-second code

[0077] A further preferred embodiment of step S3 is to increase the GeoSOT size by one level based on the GeoSOT level compared to the optimal GeoSOT level (as shown in Table 5, increasing the GeoSOT size by one level corresponds to reducing the number of geosot levels, that is, the direction of the number of geosot levels from 25 to 2 is the direction of increasing the GeoSOT size by one level), and perform mask extraction slicing, grid coding, and downsampling processing based on the grid of the increased GeoSOT level to obtain slice data containing grid coding information and the increased GeoSOT level; and obtain slice data of each GeoSOT level by increasing the GeoSOT level step by step.

[0078] In order to meet the needs of more abundant data, step S3 of this embodiment also includes the following method:

[0079] Based on the GeoSOT level, the GeoSOT size is reduced by one level compared to the optimal GeoSOT level (for example, the remote sensing image resolution is 0.01 meters, and the 22nd level of Tiandi Map corresponds to the optimal GeoSOT level of 24. Slice data of levels 23 to 2 can be obtained through mask extraction slicing, grid coding, and downsampling processing; this embodiment reduces the GeoSOT size by one level from the optimal GeoSOT level, that is, only 25 levels. In order to enrich the data, slice data of the 25-level GeoSOT level is also generated, and the 25-level slice data is obtained through mask extraction slicing, grid coding, and upsampling processing), mask extraction slicing, grid coding, and upsampling processing are performed based on the grid of the GeoSOT level reduced by one level to obtain slice data containing grid coding information and the GeoSOT level reduced by one level; the GeoSOT level is reduced step by step to obtain slice data of the remaining GeoSOT levels.

[0080] The present invention adopts a distributed file system to perform distributed storage of all slice data, and subsequent screening and rendering all adopt distributed data retrieval and rendering operations.

[0081] S4. Based on the selected sky and earth map level of the query viewport (which can obtain the current map's zoom level and viewport range), all slice data of the corresponding GeoSOT level are filtered and rendered to generate a visualization image (the visualization service informs the map of the image information to be generated while generating the image, so the image generation and loading are performed simultaneously; the map calculates the position of each image based on the position of the starting image and the GeoSOT encoding information of each image, and renders it on the map).

[0082] In some embodiments, as Figure 4 As shown, all slice data at the GeoSOT level are filtered out and the data content is extracted for standardization and color rendering. Color matching can be performed according to unique values, multiple color bands (color bands can be divided by value or by percentage) and RGB true colors to generate rendered visualization images.

[0083] like Figure 5 As shown, a high-performance visualization system for large-scale remote sensing image data includes a data acquisition module, a slice code storage system and a query viewport system. The slice code storage system has a remote sensing image sky map correspondence library, which includes a sky map level, a GeoSOT level and a level correspondence. The sky map level includes the sky map resolution, and the GeoSOT level includes the GeoSOT size. The data acquisition module is used to collect remote sensing image data and input it into the slice code storage system. The slice code storage system extracts the remote sensing image resolution corresponding to the sky map from the remote sensing image data. The query viewport system comprises a filtering module and a rendering module. The filtering module is used to select a sky and earth map level and filter all the slice data of the GeoSOT level to input into the rendering module. The rendering module uses all the slice data to render and generate a visual image.

[0084] An electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one processor to perform the steps of the high-performance visualization method for large-scale remote sensing image data of the present invention.

[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A high-performance visualization method for large-scale remote sensing image data, characterized by: The methods include: S1. Construct a remote sensing imagery sky map correspondence library including sky map levels, GeoSOT levels, and level correspondences. The sky map level includes sky map resolution, and the GeoSOT level includes GeoSOT size. The sky map resolution of the sky map level increases in sequence, while that of the GeoSOT level decreases in sequence. The level correspondence is a one-to-one correspondence between the sky map level and the GeoSOT level. The highest sky map resolution of the sky map level corresponds to the lowest GeoSOT size of the GeoSOT level, and the lowest sky map resolution of the sky map level corresponds to the highest GeoSOT size of the GeoSOT level. S2. Collect remote sensing image data, extract remote sensing image resolution corresponding to the adapted sky and earth map level and the corresponding optimal GeoSOT level, slice the remote sensing image data according to the grid of the optimal GeoSOT level and store it as optimal level slice data; S3, starting from the optimal GeoSOT level, grid slicing and downsampling are performed step by step according to the GeoSOT level to obtain slice data of each GeoSOT level in turn; S4. Based on the selected sky and earth map level of the query viewport, all slice data corresponding to the GeoSOT level are filtered and rendered to generate a visualization image.

2. The high-performance visualization method for large-scale remote sensing image data according to claim 1, characterized in that: Step S3 also includes the following method: The optimal level slice data and the slice data of each GeoSOT level are GeoSOT grid coded according to the grid.

3. The high-performance visualization method for large-scale remote sensing image data according to claim 1, characterized in that: In step S2, the sky map resolution of the sky map level is adapted according to the extracted remote sensing image resolution and the sky map level is adapted. Based on the remote sensing image sky map correspondence library, the corresponding GeoSOT level is found as the optimal GeoSOT level; mask extraction slicing and grid coding processing are performed based on the grid of the optimal GeoSOT level to obtain the optimal level slice data containing grid coding information.

4. The high-performance visualization method for large-scale remote sensing image data according to claim 1, characterized in that: In step S3, the GeoSOT size is increased by one level compared to the optimal GeoSOT level based on the GeoSOT level, and mask extraction slicing, grid coding, and downsampling processing are performed based on the grid of the increased GeoSOT level to obtain slice data containing grid coding information and the increased GeoSOT level; the GeoSOT level is increased step by step to obtain slice data of each GeoSOT level.

5. The high-performance visualization method for large-scale remote sensing image data according to claim 4, characterized in that: Step S3 also includes the following method: Based on the GeoSOT level, the GeoSOT size is reduced by one level compared to the optimal GeoSOT level. Mask extraction, slicing, grid encoding, and upsampling are performed on the grid of the reduced GeoSOT level to obtain slice data containing grid coding information and reduced GeoSOT level by one level; the GeoSOT level is gradually reduced to obtain slice data of the remaining GeoSOT levels.

6. The high-performance visualization method for large-scale remote sensing image data according to claim 5, characterized in that: A distributed file system is used to store all slice data in a distributed manner.

7. The high-performance visualization method for large-scale remote sensing image data according to claim 1, characterized in that: In step S4, all slice data at the GeoSOT level are screened and the data content is extracted for standardization and color rendering to generate a rendered visualization image.

8. A high-performance visualization system for large-scale remote sensing image data, characterized by: The system comprises a data acquisition module, a slice code storage system and a query viewport system. The slice code storage system has a remote sensing image sky map correspondence library, which includes a sky map level, a GeoSOT level and a level correspondence. The sky map level includes a sky map resolution, and the GeoSOT level includes a GeoSOT size. The sky map resolution of the sky map level increases in sequence, and that of the GeoSOT level decreases in sequence. The level correspondence is a one-to-one correspondence between the sky map level and the GeoSOT level. The highest sky map resolution of the sky map level corresponds to the lowest GeoSOT size of the GeoSOT level, and the lowest sky map resolution of the sky map level corresponds to the highest GeoSOT size of the GeoSOT level. The data acquisition module is used to collect remote sensing image data and input it into the slice coding storage system. The slice coding storage system extracts the remote sensing image resolution of the remote sensing image data corresponding to the adapted world map level and the corresponding optimal GeoSOT level, slices the remote sensing image data according to the grid of the optimal GeoSOT level and stores it as optimal level slice data; the slice coding storage system performs grid slicing and downsampling processing step by step according to the GeoSOT level starting from the optimal GeoSOT level, and obtains slice data of each GeoSOT level in turn; the query viewport system includes a screening module and a rendering module. The screening module is used to select the world map level and correspondingly screen all slice data of the GeoSOT level and input them into the rendering module. The rendering module uses all slice data to render and generate a visual image.

9. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method according to any one of claims 1 to 7.

Images