本文選題：群體動(dòng)畫 + 實(shí)時(shí)渲染��； 參考：《電子科技大學(xué)》2010年碩士論文

【摘要】： 大規(guī)模群體動(dòng)畫一直是虛擬現(xiàn)實(shí)技術(shù)研究的熱點(diǎn)之一。群體動(dòng)畫技術(shù)廣泛應(yīng)用于游戲、電影動(dòng)漫、建筑規(guī)劃以及火災(zāi)等災(zāi)難現(xiàn)場模擬與輔助指揮。通過群體行為的真實(shí)模擬,可以輔助電影動(dòng)漫產(chǎn)生真實(shí)宏大場景效果;可以在公共場所(商場、體育場館、娛樂場所等)規(guī)劃設(shè)計(jì)中模擬驗(yàn)證安全通道等設(shè)計(jì)的合理性;可以為智能交通管理(紅綠燈有效設(shè)置,主干路口的交通流量管理)提供模擬環(huán)境;也可以為災(zāi)難現(xiàn)場指揮提供指導(dǎo)和訓(xùn)練平臺。目前這個(gè)方面的研究得到了廣泛的重視,對群體動(dòng)畫的研究和創(chuàng)作工具產(chǎn)生了極大的需求,同時(shí)一套完整高效的群體動(dòng)畫設(shè)計(jì)與制作工具也將促進(jìn)影視動(dòng)畫等方面數(shù)字媒體產(chǎn)業(yè)的發(fā)展。 作為群體動(dòng)畫技術(shù)的一個(gè)關(guān)鍵技術(shù),模型簡化占據(jù)著不可替代的重要位置。要實(shí)現(xiàn)大規(guī)模的場景,必然會涉及要繪制和渲染大量的模型,而對于實(shí)時(shí)場景渲染來說,渲染速度和模型的視覺逼真度是兩個(gè)非常重要的指標(biāo)。要達(dá)到兩個(gè)指標(biāo)的和諧統(tǒng)一,那么就必須選擇一個(gè)很好的算法來實(shí)現(xiàn)渲染速度和模型顯示逼真度的折中。 對于靜態(tài)模型,現(xiàn)有技術(shù)已經(jīng)很好的實(shí)現(xiàn)了模型簡化。根據(jù)簡化算法的簡化方式可將其分為四大類:頂點(diǎn)聚類算法、增量式簡化算法、采樣算法和自適應(yīng)細(xì)分算法。在這四類模型簡化算法的基礎(chǔ)上發(fā)展起來的靜態(tài)大規(guī)模場景簡化技術(shù)也比較完善,如基于細(xì)節(jié)層次模型的大規(guī)模場景簡化技術(shù)和基于點(diǎn)采樣的大規(guī)模場景簡化技術(shù)等。本文就以上算法和技術(shù)做了具體的研究和分析,并生成了一種新的算法:一種基于點(diǎn)采樣的模型簡化算法和大規(guī)模場景簡化技術(shù)。 雖然靜態(tài)大規(guī)模場景簡化技術(shù)已經(jīng)比較完善,但動(dòng)態(tài)的大規(guī)模場景的簡化技術(shù)才剛起步,實(shí)現(xiàn)效果也不理想。針對這種情況,在提出的新算法中,加入了動(dòng)態(tài)模型的簡化新方法。實(shí)驗(yàn)結(jié)果表明:在保證場景的渲染速度和視覺真實(shí)性的條件下,能很好的實(shí)現(xiàn)30000人的實(shí)時(shí)動(dòng)畫。
[Abstract]:Mass animation has always been one of the hotspots in virtual reality technology. Group animation technology is widely used in game, movie animation, architectural planning and fire disaster scene simulation and auxiliary command. Through the real simulation of group behavior, it can assist the movie animation to produce real and grand scene effect, and can simulate and verify the rationality of the safety channel design in the planning and design of public places (shopping malls, stadiums, entertainment places, etc.); It can provide a simulated environment for intelligent traffic management (traffic lights and traffic flow management at trunk junctions), as well as a guidance and training platform for disaster scene command. At present, the research on this aspect has received extensive attention, and there is a great demand for the research and creation tools of group animation. At the same time, a set of complete and efficient group animation design and production tools will also promote the development of digital media industry. As a key technology of group animation, model simplification occupies an irreplaceable position. In order to realize large-scale scene, it is necessary to draw and render a large number of models. For real-time scene rendering, the rendering speed and the visual fidelity of the model are two very important indicators. In order to achieve the harmony and unity of the two indexes, we must choose a good algorithm to achieve the compromise between rendering speed and model display fidelity. For static models, the existing techniques have achieved model simplification well. According to the simplified algorithm, it can be divided into four categories: vertex clustering algorithm, incremental simplification algorithm, sampling algorithm and adaptive subdivision algorithm. The static large-scale scene simplification technology based on these four models simplification algorithms is also relatively perfect, such as the large-scale scene simplification technology based on detail hierarchy model and the large-scale scene simplification technology based on point sampling. In this paper, the above algorithms and techniques are studied and analyzed in detail, and a new algorithm is generated: a model simplification algorithm based on point sampling and a large-scale scene simplification technique. Although the static large-scale scene simplification technology has been relatively perfect, but the dynamic large-scale scene simplification technology has just started, the realization effect is not ideal. In order to solve this problem, a new simplified method of dynamic model is added to the proposed algorithm. The experimental results show that under the condition of guaranteeing the rendering speed and visual authenticity of the scene, the real-time animation of 30000 people can be realized very well.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2010
【分類號】：TP391.41

【引證文獻(xiàn)】

相關(guān)碩士學(xué)位論文 前1條

1 高楊;人車混合環(huán)境智能控制研究與實(shí)現(xiàn)[D];電子科技大學(xué);2012年

，

本文編號：1821723