Particular attention is paid to the use of the color channels in estimating the motion. The main premise is that three different intensity channels conveying the same motion information and yielding more accurate motion estimates.
提高视频图像序列的空间分辨率主要取决于运动估计算子的精度，重点放在利用(R,G,B)颜色通道进行运动估计，其主要前提是三个不同的颜色通道传递着相同的运动信息并且能够产生更加精确的运动估计。
参考来源 - 视频图像超分辨率复原·2,447,543篇论文数据，部分数据来源于NoteExpress
Same-parity fields are used to improve the accuracy of motion estimation. It estimates the motion by comparing the sum of absolute block difference(SAD) between 4 fields with the threshold of motion.
运动估计通过比较同极性场的像素块绝对值差和（SAD）与运动阈值实现运动估计，并且利用前一场的运动信息进行鲁棒性处理，提高运动检测的精度。
Two methods of initial estimates are proposed: (1) using real time image subtraction scheme and a micro motion compensation mechanism to take zero constantly as initial estimates;
对相关迭代算法中的初值估计问题进行了研究，提出两种初值估计方法：（1 ）利用“实时相减”和“精密调节”相结合的方法而获得零初值；
应用推荐