The 3D terrain rendering is widely used in many industries, such as the military, agriculture, transportation, and smart cities. The processing method for image data mainly adopts spatial meshing and establishes tile data at different levels. In order to improve the resolution of the 3D terrain, the number of sliced tiles is increasing exponentially, and the accessing, indexing, and rendering of massive tile data are technical problems that need to be overcome. In this paper, a real-time rendering method for layered image data based on TMS (tile map senice) was proposed to solve above problems and to address the severe chromatic aberration at the junction of different levels of images. In this method, service objects with different image layers were constructed, and the flexible control and management of the hierarchy, order, and position of digital twin city hierarchical images were realized. Meanwhile, three queues were stored for high, medium, and low relative priority of the loaded tiles in the viewport to speed up the scheduling and rendering of tile data. To accelerate the acquisition of image tile index, the screen pixel size was calculated according to the tile matrix range in the viewport. The final level of the image was determined by comparing it with the level of the current file resource after obtaining the preliminary level of the image tile. Then, the index of the image tile in the viewport was calculated, and the row and column number were calculated according to the latitude, longitude, and level to form the request path of the image resource. Finally, the two layers were blended sequentially from the lowest layer to obtain the correct blending result according to the size of the image hierarchy. In the paper, a set of 3D terrain rendering experiments were carried out on near, medium, far, and coastal multi-layer image data in practical applications. The results showed that the NVIDIA 2070 graphics card can achieve high real-time performance of 300 FPS and high-quality rendering. Therefore, the proposed method can flexibly render large-scale and multi-level image data to construct a 3D terrain of the digital twin cities. © 2023 Editorial Department of Journal of Sichuan University. All rights reserved.
