【摘要】 基于蒙特卡罗光线追踪的全局光照绘制方法,在图像真实感绘制领域起到了关键的作用。但是由于噪声在图像中分布的不均匀性,对每个像素采用相同的采样点不是最优的采样方法,因此,论文提出一种基于像素块方差的自适应采样方法。该方法通过像素块的颜色通道方差值与分裂阀值的比较确定该像素块是否需要进一步分裂。随着采样点的增加以及算法的迭代,最终会将整个图像分裂成若干个子像素块,方差较大的区域给与更多的采样点,方差较小的区域分配较少的采样点,从而提高绘制效率,优化绘制效果。更多还原

【Abstract】 Monte Carlo-based global illumination method plays a key role in the field of image realistic rendering. However,due to the uneven distribution of noise in images,it is not the best sampling method to use the same sampling points for each pixel.Therefore,an adaptive sampling method based on pixel block variance is proposed in this paper. This method determines whether the pixel block needs to be further split by comparing the variance value of the color channel of the pixel block with the split threshold. With the increase of the sampling point and the iteration of the algorithm,the whole image will eventually be split into a number of sub pixel blocks,and the regions with larger variance will be given more multiple sampling points,but the regions with smaller variance reduce sampling points. Therefore,the efficiency of drawing is improved and the effect of drawing is optimized.更多还原