CAJ | 학술논문

针对海量数据难以实时渲染的问题,提出基于视点的大规模点云数据实时预测调度策略--"5-5-5"动态调度策略.对整个三维点云进行"分区-子块化"处理,经过多分辨率压缩后保存到服务器端的数据库中.客户端根据视点的变化和菱形调度规则对子块数据进行内外存动态调度并实时绘制,同时向服务器发送传输新数据的请求.实验证明,该算法能有效解决漫游过程中的"抖动"现象,减少单次向服务器请求传输的数据量,实现漫游显示的连续性和平滑性.
침대해량수거난이실시선염적문제,제출기우시점적대규모점운수거실시예측조도책략--"5-5-5"동태조도책략.대정개삼유점운진행"분구-자괴화"처리,경과다분변솔압축후보존도복무기단적수거고중.객호단근거시점적변화화릉형조도규칙대자괴수거진행내외존동태조도병실시회제,동시향복무기발송전수신수거적청구.실험증명,해산법능유효해결만유과정중적"두동"현상,감소단차향복무기청구전수적수거량,실현만유현시적련속성화평활성.
To address the problem of rendering massive data in real time difficultly, a real-time predictive schedule strategy based on viewpoint for large-scale point-cloud data is proposed, that is, "5-5-5" dynamic schedule strategy. The entire three-dimensional point-cloud is processed into "district-subblocks". And these subblocks are saved in the server database after multi-resolution compression. According to the changes of the view parameters and diamond-shaped schedule, subblock data can be scheduled dynamically and rendered in real time. The server is asked for transmitting the data expected to the client, Experimental results show that the algorithm effectively alleviates jitters and reduces one-time transmission payload from server, with good achievement of continuity and smoothness in process of roaming.