A new satellite image display method

Abstract

The application relates to the field of satellite image display methods, in particular to a novel satellite image display method, which comprises the following steps: constructing a viewport state vector containing a viewport position, a level and a speed; constructing a retention value evaluation model based on a spatial Euclidean distance and a level span, scoring the values of the cached tiles to determine the elimination order; constructing a prefetch priority model based on the dot product of a motion speed vector and a displacement vector, and performing directional preloading on the peripheral tiles of the viewport; and dynamically executing cache replacement and data prefetching through resource monitoring closed loop. The application can improve the cache hit rate, eliminate the zoom white screen phenomenon, and realize a smooth satellite image roaming experience under the condition of limited hardware.

Description

Technical Field

[0001] This invention relates to the field of satellite image display methods, and more particularly to a novel satellite image display method. Background Technology

[0002] With the rapid development of remote sensing technology, the resolution of satellite imagery data is constantly improving, and the data volume has reached TB or even PB levels. In applications such as aviation vocational education, flight simulation training, and UAV ground stations, it is necessary to load and display this massive amount of imagery data in real time on the operating terminal. However, the hardware performance of such terminal devices is often limited, for example, insufficient video memory bandwidth or limited memory capacity. Coupled with the uncertainty of network transmission, this poses a significant challenge to the real-time display of data.

[0003] Existing technologies typically employ a pyramid model combined with tile technology to organize data, and use the Least Recently Used (LRU) algorithm to manage memory caching. The core logic of the LRU algorithm is to evict the data that has not been accessed for the longest time. However, this strategy based on a single time dimension has significant drawbacks in the specific scenario of satellite image display: First, it lacks consideration of geospatial correlation, easily leading to the accidental deletion of data that is spatially adjacent to the current viewport but has been temporarily moved out, causing stuttering when the user revisits the image; second, it ignores the vertical correlation between pyramid levels, easily evicting parent data that is about to become the background during rapid zooming, causing a white screen phenomenon.

[0004] Furthermore, existing caching strategies lack the ability to proactively predict motion trends. Most current technologies are passively responsive, meaning they only begin loading data after a page fault occurs, failing to proactively intervene based on the speed and direction vectors of user operations. In high-speed cruising or rapid dragging scenarios, data loading often lags behind viewport movement, resulting in black or blank areas at the edges of the screen, severely impacting visual smoothness and user experience. Summary of the Invention

[0005] To address the technical problem that the LRU algorithm in the prior art lacks spatial and hierarchical considerations and cannot actively adapt to motion trends, resulting in display stuttering and blank screens, this invention provides a novel satellite image display method.

[0006] This invention provides a novel satellite image display method, employing the following technical solution: A novel satellite image display method includes the following steps: The process initializes the tile pyramid index structure and buffer pool, obtains the viewport parameters in the geographic projection coordinate system in real time, and calculates the viewport state vector at the current moment based on the parameters. The viewport state vector includes the viewport center coordinates, real-time zoom level, and viewport movement speed. When memory usage exceeds a preset high watermark threshold, the process traverses the candidate tiles in the buffer pool. Based on the Euclidean distance between the tile center and the viewport center and the difference between the tile level and the real-time zoom level, the process calculates the retention value score for each candidate tile and sorts and cleans the candidate tiles according to the retention value score. Utilizing idle I / O resources, the process calculates the prefetch priority of unloaded tiles within the viewport expansion area based on the direction and magnitude of the viewport movement speed. The prefetch priority is positively correlated with the consistency of the tile relative to the movement direction and negatively correlated with the distance from the tile to the viewport. During the rendering loop, the process monitors the memory status in real time, performs cache replacement based on the retention value score, generates a loading queue based on the prefetch priority, performs texture mapping on ready tiles, and performs blur rendering on parent tiles for unready areas.

[0007] To address the issue that the LRU (Least Recently Used) algorithm, commonly used in existing technologies, manages the cache based solely on the time dimension, which can easily lead to the accidental deletion of spatially adjacent or parent background data, causing stuttering and blank screens, this invention proposes an active scheduling strategy based on multi-dimensional viewport states. By comprehensively considering the spatial Euclidean distance between tiles and the viewport, as well as hierarchical differences, it calculates retention value, intelligently retaining the data most important to the current view. Simultaneously, it utilizes the magnitude and direction of viewport movement speed to calculate prefetch priority, achieving proactive preloading in accordance with movement trends. This improves cache hit accuracy, effectively avoids visual gaps during rapid dragging or zooming, and enhances the smoothness of user operations.

[0008] Preferably, the expression for the viewport state vector is:

[0009] In the formula, For the current moment The viewport state vector, and This represents the absolute coordinates of the current viewport center point in the Mercator projection coordinate system. For the current real-time scaling level, and This represents the current horizontal and vertical movement speed of the viewport on the projection plane.

[0010] By explicitly defining a viewport state vector that includes the absolute coordinates of the viewport center, the real-time zoom level, and the horizontal and vertical movement speeds, a mathematical basis for accurately describing the current observation state is provided. Compared with existing technologies that only focus on static position, the introduction of real-time zoom level and movement speed parameters enables the system to dynamically perceive the user's operational intentions, thereby providing accurate decision-making basis for subsequent cache replacement and data prefetching algorithms, ensuring that the data scheduling logic is consistent with the user's actual operational behavior.

[0011] Preferably, the retention value score for each candidate tile is calculated using the following formula:

[0012] In the formula, For tiles Retention value score, and These are the spatial weighting coefficient and the hierarchical weighting coefficient, respectively. For tiles The Euclidean distance between the center point of the viewport and the current viewport center point The field of view radius of the current viewport. Let be the spatial distance regularization constant. For tiles The discrete level to which it belongs. For the current real-time scaling level, This is the hierarchical deviation correction constant.

[0013] By adjusting the spatial weight coefficient and the hierarchical weight coefficient, the retention priority of tiles is quantified. This formula not only considers the planar distance, but also prevents the parent tile from being eliminated prematurely through hierarchical deviation correction. Compared with the traditional algorithm that blindly eliminates old data, this formula ensures that key tiles that are spatially adjacent to the viewport or hierarchically serve as background references can obtain higher retention scores, thereby maximizing the continuity and integrity of the image when memory is tight.

[0014] Preferred Euclidean distance The acquisition process includes: Calculate the center coordinates of the tile based on its row and column index. Calculate the square of the difference between the x-coordinate of the tile's center point and the x-coordinate of the current viewport's center point, and the square of the difference between the y-coordinate of the tile's center point and the y-coordinate of the current viewport's center point. Add the two squared differences together and take the arithmetic square root to obtain the Euclidean distance.

[0015] The actual distance between the tile center and the viewport center is calculated by inverting the coordinates using row and column numbers and combining this with the Pythagorean theorem. This ensures the accuracy of the core parameter of spatial distance in the retention value assessment model, enabling the cache eviction strategy to be strictly executed based on geographical proximity, thus guaranteeing the availability of data around the field of view.

[0016] Preferably, the prefetch priority of unloaded tiles within the viewport expansion area is calculated using the following formula:

[0017] In the formula, For tiles Prefetch priority, The current motion velocity vector of the viewport. To point from the center of the viewport to the tile The displacement vector at the center, displacement vector The model, Let be the velocity direction regularization constant. This is the distance attenuation correction constant.

[0018] By calculating the relationship between the motion velocity vector and the tile displacement vector, the urgency of loading is quantified. This formula enables the system to distinguish tiles from different orientations, giving higher loading weights to tiles facing the direction of motion. Compared to traditional uniform preloading around the perimeter or passive loading, this improves the targeting and effectiveness of data loading.

[0019] Preferably, when calculating the prefetch priority, the method further includes: calculating the motion velocity vector. With displacement vector The dot product is used to determine whether the tile is located behind or to the side rear of the direction of movement when the dot product value is less than or equal to zero, and the prefetch priority of the tile is adjusted accordingly. Set to zero to skip preloading.

[0020] By calculating the dot product of the velocity vector and the displacement vector to determine directional consistency, it can intelligently identify and filter out invalid tile requests located behind or to the side of the direction of movement. In high-speed cruising or rapid dragging scenarios, it effectively avoids wasting bandwidth and I / O resources on areas that the user has already browsed or is about to leave, ensuring that limited system resources are concentrated on loading the area that the user is about to look at, thereby significantly reducing loading latency.

[0021] Preferably, cache replacement is performed based on retention value score, including: If memory usage is detected to be greater than the high watermark threshold, calculate the retention value score of all cached non-display area tiles; sort all candidate tiles according to their retention value scores from smallest to largest; release the memory occupied by the tiles starting from the head of the sorted list until the total memory usage is less than the low watermark threshold.

[0022] By sorting non-display area tiles by retention value and cleaning them from low to high until memory usage falls back to a safe level, memory overflow is prevented, and performance jitter caused by frequent garbage collection is avoided. This ensures that only data that contributes the least to the current display is released, maximizing cache utilization while maintaining system stability.

[0023] Preferably, the loading queue is generated based on the prefetch priority, including: traversing the list of tiles to be loaded outside the viewport and calculating the prefetch priority of each tile; removing tiles with a prefetch priority less than or equal to zero and sorting the remaining tiles in descending order of prefetch priority; and selecting the top N tiles in the sorting to add to the asynchronous download and decoding queue.

[0024] By removing reverse tiles and selecting the N highest priority tiles into the queue, the congestion problem of the network or decoding thread is solved. This ensures that in browsing satellite imagery with huge amounts of data, the decoder and downloader always handle the most urgent tasks, eliminating the long-tail delay caused by queue backlog that prevents high-priority data from being displayed in a timely manner.

[0025] Preferably, the process of obtaining the horizontal and vertical movement speed includes: recording the viewport center coordinates of the previous frame and the viewport center coordinates of the current frame; calculating the coordinate difference between the viewport center coordinates of the current frame and the previous frame, and dividing the coordinate difference by the time interval between the two frames to obtain the instantaneous movement speed; performing a moving average filter on the instantaneous movement speed to eliminate operation jitter noise, and obtaining the final horizontal and vertical movement speed.

[0026] A moving average filter is introduced to obtain the final movement speed, effectively eliminating speed abrupt changes caused by slight hand tremors or sampling noise from the input device. This provides a smooth and stable speed input for the prefetching algorithm, preventing frequent jumps or misjudgments in the preloading queue due to drastic fluctuations in speed parameters, thus enhancing the robustness of the prediction algorithm.

[0027] Preferably, the initialization of the tile pyramid index structure and buffer pool includes: using a quadtree structure to perform multi-resolution slicing of global satellite imagery and defining the basic pixel size of the tiles; allocating a fixed-capacity physical memory region in memory as a tile buffer pool; and constructing a bidirectional mapping table between the unique tile ID containing the tile level, row number, and column number and the memory address, so as to support indexing the memory address by the tile ID and looking up the tile attribute by the memory address.

[0028] By employing a quadtree structure combined with a fixed-capacity physical memory buffer pool and establishing a bidirectional mapping table, the problems of indexing efficiency and memory fragmentation for massive tile data are solved.

[0029] The present invention has the following technical effects: This invention utilizes state vectors containing position, level, and velocity to construct a tile retention value model based on spatial Euclidean distance and level difference, achieving accurate cache replacement. Simultaneously, it uses the dot product of motion velocity vector and displacement vector to calculate prefetch priority, achieving directional active preloading in accordance with the motion direction. This method effectively solves the problems of image browsing lag and white screen caused by the lack of spatial correlation consideration and motion trend prediction. Attached Figure Description

[0030] Figure 1 This is a flowchart of a novel satellite image display method according to the present invention.

[0031] Figure 2 This is a comparison test chart of cache hit rate during the satellite image browsing process of this invention.

[0032] Figure 3 This is a comparison chart of the average time distribution for loading tile data according to the present invention.

[0033] Figure 4 This is a visualization comparing the validity of memory data in a high-speed roaming scenario according to the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] This invention discloses a novel satellite image display method, referring to... Figure 1 This includes steps S101-S104: S101: Initialize the tile pyramid index structure and buffer pool, obtain the parameters of the viewport in the geographic projection coordinate system in real time, and calculate the viewport state vector at the current moment based on the parameters. The viewport state vector includes the viewport center coordinates, real-time zoom level, and viewport movement speed.

[0036] In its implementation, the system first employs a quadtree structure to slice global satellite imagery at multiple resolutions, defining the basic pixel size of each tile, such as 256×256 pixels. Simultaneously, a fixed-capacity physical memory region is allocated as a tile buffer pool, and a bidirectional mapping table is constructed. A unique tile ID, containing the level (L), row number (Row), and column number (Col), can be directly indexed to a memory address. Furthermore, high-water mark thresholds (e.g., 80%) and low-water mark thresholds (e.g., 70%) are set as boundary conditions to trigger cache replacement.

[0037] Furthermore, using the rendering engine's frame callback function as the trigger point, for example, triggering once every 16.6ms, the parameters of the current viewport in the geographic projection coordinate system are obtained in real time, defining the current moment. The viewport state vector is :

[0038] in, and The absolute coordinates of the current viewport center point in the Mercator projection coordinate system are obtained by reading the latitude and longitude of the camera focus and transforming it through projection. For the current real-time scaling level, retain two decimal places, such as 14.35 levels, for accurate calculation of level distances; and The horizontal and vertical movement velocities of the current viewport on the projection plane are calculated using a differential approximation method: the previous frame time is recorded. center coordinates and time interval ,but , The speed is then filtered using a moving average to eliminate noise caused by hand tremors.

[0039] For example, suppose the viewport center coordinates of the previous frame are... The current frame coordinates are The time interval is ,but Pixels per second, indicating that the user is moving quickly to the right.

[0040] In this way, by constructing a multi-dimensional state vector that includes viewport position, hierarchy, and speed, it is possible to accurately perceive the user's operational intent and the physical state of the current viewport, providing a precise basis for subsequent data scheduling and fundamentally changing the status quo of blind management by traditional algorithms.

[0041] S102: When memory usage exceeds the high water level threshold, traverse the candidate tiles in the buffer pool, calculate the retention value score of each candidate tile based on the Euclidean distance between the tile center and the viewport center and the difference between the tile level and the real-time scaling level, and sort and clean the candidate tiles according to the retention value score.

[0042] In practice, when memory usage exceeds the high-water mark threshold, a cache replacement process is triggered. At this point, all candidate tiles outside the currently displayed area in the buffer pool are traversed, their retention value scores are calculated, and any cached tile is defined. The retention value score is The calculation formula is as follows:

[0043] In the formula, For tiles The overall retention value is considered, with higher values ​​indicating higher priority. These are the spatial weighting coefficient and the hierarchical weighting coefficient, with a range of values. And the sum is 1. For tiles The Euclidean distance between the center point of the viewport and the current viewport center point The field of view radius of the current viewport. Let be the spatial distance regularization constant. For tiles The discrete level to which it belongs. This is the current real-time scaling level. This is the hierarchical deviation correction constant.

[0044] To facilitate understanding, a simple calculation example is given below: Assuming the parameters are set as follows: .

[0045] Current viewport status: center point Real-time level .

[0046] Tile A (nearby tiles): center point hierarchical .

[0047] Calculation process: Spatial distance Spatial item: Contribution value Hierarchical difference: Hierarchical items: Contribution value Total score .

[0048] Tile B (far end tile): center point hierarchical .

[0049] Calculation process: Spatial distance Spatial item: The contribution is minimal, and the hierarchical item is the same as above. Total score .

[0050] Tile A has a much higher score than tile B, so during memory cleanup, tile B will be prioritized for elimination, while tile A will be retained due to spatial proximity.

[0051] Thus, by comprehensively considering spatial Euclidean distance and hierarchical span, the evaluation model can intelligently identify and retain key tiles that are spatially adjacent or hierarchically related, effectively avoiding I / O oscillations and scaling white screens caused by the accidental deletion of high-value data.

[0052] S103: Utilizing idle I / O resources, calculate the prefetch priority of unloaded tiles within the viewport expansion area based on the direction and magnitude of the viewport movement speed. The prefetch priority is positively correlated with the consistency of the tile relative to the movement direction and negatively correlated with the distance from the tile to the viewport.

[0053] In practice, while clearing memory, preloading is performed using idle I / O resources. This step calculates the prefetch priority of the tiles to be loaded outside the viewport to achieve directional data preparation in accordance with the direction of movement, defining any unloaded tile within the viewport expansion area. The prefetch priority is The calculation formula is as follows:

[0054] In the formula, The current motion velocity vector of the viewport , To point from the center of the viewport to the tile Displacement vector at center , For the dot product operation of two vectors, displacement vector The unit is meters. Here is the velocity direction regularization constant, in meters. This is the distance attenuation correction constant, exemplified by a value of 1.

[0055] It should be noted that when the dot product result At that time, it was explained that the tile Located behind or to the side of the direction of movement, or in a stationary state. At this time, directly order... No preloading is performed.

[0056] The following is a calculation example: Assuming viewport center Moving due east, velocity vector .

[0057] Tile C (directly in front): Center displacement vector ,mold dot product (Greater than 0, same direction); Substitute into the formula (let) ): .

[0058] Tile D (directly behind): Center displacement vector dot product Since the dot product is less than 0, it is determined to be a deviation from the direction, so let... .

[0059] Therefore, the system will prioritize loading the tile C directly in front, while completely ignoring the tile D behind.

[0060] Thus, the directional preloading strategy based on velocity vector dot product achieves precise data preparation, eliminates invalid requests behind the direction of motion, greatly saves bandwidth resources, and ensures visual continuity in high-speed motion scenarios.

[0061] S104: Monitor memory status in real time during the rendering loop, perform cache replacement based on the retention value score, generate a loading queue based on the prefetch priority, perform texture mapping on ready tiles, and call the parent tile for blur rendering of unready areas.

[0062] In practice, the above evaluation and calculation logic is embedded into an independent thread of the main rendering loop to form a closed-loop control, and the heap memory usage of video memory and main memory is detected every frame.

[0063] If memory usage exceeds the high watermark threshold, calculate the retention value score for all cached non-display area tiles. ;Place list by Sort by size from smallest to largest; start releasing tile memory from the head of the queue until the total memory usage is less than the low watermark threshold.

[0064] If the current I / O thread pool has idle resources, iterate through the list of unloaded tiles outside the viewport; calculate the load value for each tile. Remove The remaining tiles are arranged according to the tiles. Sort the tiles from largest to smallest; add the top-N tiles to the asynchronous download and decoding queue.

[0065] The rendering thread only reads tiles with a "ready" status from the buffer pool for rendering. For areas that are not ready, it automatically searches upwards to find the parent tile texture and performs magnified and blurred rendering as a temporary placeholder.

[0066] like Figure 2 As shown, in satellite image browsing involving continuous operations such as panning and zooming, the cache hit rate of the method of this invention is high, significantly better than existing technologies, especially in the rapid zooming and panning range, where the hit rate advantage of the method of this invention is even more obvious. Figure 3 As shown, the method of the present invention has a lower average tile loading time and eliminates long-tail high latency, indicating that the method of the present invention can maintain stable loading performance under different operating scenarios. Figure 4 As shown, in high-speed roaming scenarios, the data retained by the method of this invention tightly surrounds the viewport and extends in the direction of motion, while the data retained by existing methods exhibits a fragmented distribution. This demonstrates that the method of this invention can concentrate limited memory resources on high-value areas, improving the spatial efficiency of memory data.

[0067] In this way, by dynamically executing cache replacement and data prefetching through closed-loop resource monitoring, the system can always maintain the availability of core data when memory resources are scarce. Furthermore, by using the blur rendering of the parent tile as a fallback, visual voids are completely eliminated, providing a smooth and natural operating experience.

[0068] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A novel satellite image display method, characterized in that, Including the following steps: Initialize the tile pyramid index structure and buffer pool, obtain the parameters of the viewport in the geographic projection coordinate system in real time, and calculate the viewport state vector at the current moment based on the parameters. The viewport state vector includes the viewport center coordinates, real-time zoom level, and viewport movement speed. When memory usage exceeds the preset high water level threshold, the candidate tiles in the buffer pool are traversed. Based on the Euclidean distance between the tile center and the viewport center and the difference between the tile level and the real-time scaling level, the retention value score of each candidate tile is calculated, and the candidate tiles are sorted and cleaned according to the retention value score. Utilizing idle I / O resources, the prefetch priority of unloaded tiles within the viewport expansion area is calculated based on the direction and magnitude of the viewport movement speed. The prefetch priority is positively correlated with the consistency of the tile relative to the movement direction and negatively correlated with the distance of the tile from the viewport. During the rendering loop, the memory status is monitored in real time, cache replacement is performed based on the retention value score, and a loading queue is generated based on the prefetch priority. Texture mapping is performed on ready tiles, and parent tiles are called for blur rendering of unready areas.

2. The novel satellite image display method according to claim 1, characterized in that, The expression for the viewport state vector is: In the formula, For the current moment The viewport state vector, and This represents the absolute coordinates of the current viewport center point in the Mercator projection coordinate system. For the current real-time scaling level, and This represents the current horizontal and vertical movement speed of the viewport on the projection plane.

3. The novel satellite image display method according to claim 1, characterized in that, The retention value score for each candidate tile is calculated using the following formula: In the formula, For tiles Retention value score, and These are the spatial weighting coefficient and the hierarchical weighting coefficient, respectively. For tiles The Euclidean distance between the center point of the viewport and the current viewport center point The field of view radius of the current viewport. Let be the spatial distance regularization constant. For tiles The discrete level to which it belongs. For the current real-time scaling level, This is the hierarchical deviation correction constant.

4. The novel satellite image display method according to claim 3, characterized in that, Euclidean distance The acquisition process includes: The center coordinates of the tile are calculated by reversing the row and column indexes of the tile. The square of the difference between the x-coordinate of the tile center point and the x-coordinate of the current viewport center point, and the square of the difference between the y-coordinate of the tile center point and the y-coordinate of the current viewport center point are calculated. The two squared differences are added together and the arithmetic square root is taken to obtain the Euclidean distance.

5. A novel satellite image display method according to claim 1, characterized in that, The prefetch priority of unloaded tiles within the viewport expansion area is calculated using the following formula: In the formula, For tiles Prefetch priority, The current motion velocity vector of the viewport. To point from the center of the viewport to the tile The displacement vector at the center, displacement vector The model, Let be the velocity direction regularization constant. This is the distance attenuation correction constant.

6. A novel satellite image display method according to claim 5, characterized in that, The calculation of prefetch priority also includes: calculating the motion velocity vector. With displacement vector The dot product is used to determine whether the tile is located behind or to the side rear of the direction of movement when the dot product value is less than or equal to zero, and the prefetch priority of the tile is adjusted accordingly. Set to zero to skip preloading.

7. The novel satellite image display method according to claim 1, characterized in that, Cache replacement is performed based on retention value score, including: If memory usage is detected to be greater than the high watermark threshold, calculate the retention value score of all cached non-display area tiles; sort all candidate tiles according to their retention value scores from smallest to largest; release the memory occupied by the tiles starting from the head of the sorted list until the total memory usage is less than the low watermark threshold.

8. A novel satellite image display method according to claim 1, characterized in that, The loading queue is generated based on the prefetch priority, including: traversing the list of tiles to be loaded outside the viewport and calculating the prefetch priority of each tile; removing tiles with a prefetch priority less than or equal to zero and sorting the remaining tiles in descending order of prefetch priority; and selecting the top N tiles in the sorting to add to the asynchronous download and decoding queue.

9. A novel satellite image display method according to claim 2, characterized in that, The process of obtaining horizontal and vertical movement speed includes: recording the viewport center coordinates of the previous frame and the current frame; calculating the coordinate difference between the viewport center coordinates of the current frame and the previous frame, and dividing the coordinate difference by the time interval between the two frames to obtain the instantaneous movement speed; performing moving average filtering on the instantaneous movement speed to eliminate operation jitter noise, and obtaining the final horizontal and vertical movement speed.

10. A novel satellite image display method according to claim 1, characterized in that, Initialize the tile pyramid index structure and buffer pool, including: using a quadtree structure to slice global satellite imagery at multiple resolutions and defining the basic pixel size of the tiles; allocating a fixed-capacity physical memory area in memory as a tile buffer pool; and constructing a bidirectional mapping table between the unique tile ID containing the tile level, row number, and column number and the memory address, to support indexing the memory address by the tile ID and looking up the tile attribute by the memory address.

Images