本文關鍵詞： 網(wǎng)絡化應用 地形繪制 實時傳輸 地形可視化　出處：《南京師范大學》2013年碩士論文　論文類型：學位論文

【摘要】：SAAS (Software as a services,軟件服務化)是21世紀新興起的一種軟件應用模式,在SAAS模式中,用戶僅需通過瀏覽器就可實現(xiàn)軟件應用。應用程序網(wǎng)絡化是實現(xiàn)SAAS的前提,實現(xiàn)單機高效地形可視化算法的網(wǎng)絡化應用,對于基于SAAS模式的數(shù)字城市建設具有重要意義。相對于網(wǎng)絡二維地圖,網(wǎng)絡三維地形場景具有符號簡單、生動逼真特點,但基于網(wǎng)絡的三維地形場景可視化研究,受到地形數(shù)據(jù)量大、地形傳輸耗時、客戶端插件等問題的困擾,軟件服務化發(fā)展較慢。本文針對上述研究難點,做了如下幾個方面的研究： 1)研究了目前主流高效單機版地形繪制算法——GeoClipmap,在GeoClipmap算法中,每一層LOD (Level of Detail)的數(shù)據(jù)量是可控且有規(guī)律的,具有渲染幀率穩(wěn)定的優(yōu)點。本文根據(jù)GeoClipmap的規(guī)則網(wǎng)格嵌套特性,設計了新型的適用于網(wǎng)絡傳輸?shù)腉eoClipmap坐標系以及裁剪機制,利用該坐標系及裁剪機制,根據(jù)Clipmap原點,客戶端能夠快速重構出地形網(wǎng)格的X、Y坐標,相對于傳統(tǒng)傳輸完整地形頂點數(shù)據(jù)結構的方法,本文方法達到了8倍的壓縮率。 2)對于地形模型的實時傳輸,本文從底層擴展設計了WebSocket數(shù)據(jù)幀結構,并采用WebSocket進行地形網(wǎng)格數(shù)據(jù)實時傳輸。通過擴展數(shù)據(jù)幀結構,實現(xiàn)了以Clipmap層為單位的地形數(shù)據(jù)實時傳輸。另外充分結合WebSocket協(xié)議小數(shù)據(jù)包報頭特點,可以減少模型傳輸時間和增加3D模型可傳輸量。 3)對于客戶端地形繪制,本文在免安裝插件情況下,設計了一種構建三角條帶的新方法,地形網(wǎng)格數(shù)據(jù)傳輸至客戶端后,通過在每一條三角條帶頭尾添加退化三角形,將方形網(wǎng)格轉換為一條三角條帶進行繪制,進一步提高了地形繪制速度。 通過上述三方面研究,本文設計實現(xiàn)了面向網(wǎng)絡的免插件地形傳輸及可視化原型系統(tǒng),驗證了GeoClipmap算法網(wǎng)絡化應用的可行性,本原型系統(tǒng)中,每一層255×255地形網(wǎng)格的數(shù)據(jù)傳輸量由0.49M減小至0.06M,數(shù)據(jù)量減少87.8%,相對于GeoClipmap算法,數(shù)據(jù)量減少51%,大大提高了客戶端地形可視化速度。
[Abstract]:In 21th century, SAAS Software as a service is a new software application mode. In SAAS mode, users can realize software application only through browser. Application networking is the prerequisite of SAAS. It is very important for the construction of digital city based on SAAS mode to realize the network application of high efficiency terrain visualization algorithm on single computer. Compared with the two-dimensional map of network, the network 3D terrain scene has simple symbol, vivid and lifelike features. However, the research of 3D terrain scene visualization based on network is beset by the problems of large amount of terrain data, time consuming of terrain transmission, client plug-in and so on, so the development of software service is slow. The following studies have been done:. 1) GeoClipmap-the main effective single-machine topographic rendering algorithm is studied. In the GeoClipmap algorithm, the data of each layer of LOD level of detail is controllable and regular, which has the advantage of stable frame rate rendering. According to the regular mesh nesting characteristic of GeoClipmap, A new GeoClipmap coordinate system and its clipping mechanism suitable for network transmission are designed. By using the coordinate system and clipping mechanism, the client can quickly reconstruct the XY coordinate of the terrain grid according to the origin of Clipmap. Compared with the traditional method of transmitting complete topographic vertex data structure, the compression ratio of this method is up to 8 times. 2) for the real-time transmission of terrain model, this paper designs the WebSocket data frame structure from the bottom layer extension, and uses WebSocket to carry out the real-time transmission of terrain grid data. The real-time transmission of terrain data based on Clipmap layer is realized. In addition, combined with the characteristics of small packet header of WebSocket protocol, it can reduce the transmission time of the model and increase the transmissible amount of 3D model. 3) for client topographic rendering, this paper designs a new method of constructing triangulation strip without installing plug-in. After the topographic grid data is transmitted to the client, a degenerate triangle is added at the beginning and end of each triangulation strip. The square mesh is transformed into a triangular strip for rendering, which further improves the speed of terrain rendering. Through the above three aspects of research, this paper designs and implements a network-oriented terrain transmission and visualization prototype system, and verifies the feasibility of the network application of GeoClipmap algorithm. In this prototype system, The amount of data transmission in each layer of 255 脳 255 terrain grid is reduced from 0.49m to 0.06m, and the amount of data is reduced by 87.8 meters. Compared with GeoClipmap algorithm, the amount of data is reduced by 51m, which greatly improves the speed of terrain visualization on client side.
【學位授予單位】：南京師范大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：P284

【引證文獻】

相關博士學位論文 前1條

1 張俊峰;大規(guī)模地形實時動態(tài)多分辨率顯示關鍵算法研究[D];武漢大學;2011年

，

本文編號：1534910