基于卡尔曼预测器的数据预取方法

摘要

大规模复杂场景绘制,地形数据量庞大,需动态调度地形数据；但是由于磁盘传输的延迟,会出现绘制时的卡帧现象.为了解决上述问题,通过研究卡尔曼滤波器,将其与固定间隔的预取方法相结合,提出了采用卡尔曼预测器的数据预取方法.上述方法通过卡尔曼预测器预测视点位置,判断视域范围,进行数据预取.可显著提高视点预测的准确性,减少下一时刻需要预取的数据量,提高绘制效率.%The data of large - scale and complicated scene rendering are so large that requiring dynamic scheduling , but the delay of data transmission between hard disk and memory makes the rendering not fluently. To address this issue, a new Kalman predictor based data prefetching method was proposed, which combine fixed intervals data prefetching method with Kalman predictor. This method chooses the needed data based on the viewpoint position predicted by Kalman predictor, which can reduce the amount of data and improve the rendering efficiency remarkably.