KLT has been designed to be easy to use. It should take no more than a few minutes to learn how to select and track features, and store and use the results. For those who wish to maximize performance, however, the program allows the flexibility of tw

为了实时地稳定摄像系统的输出视频,提出了一种基于特征匹配的实时电子稳像算法。将改进 的Kanade2Lucas2Tomasi (KLT) 算法用于特征提取,并提出一种新的基于灰度投影均值的特征匹配算法。为了保 证稳像结果的鲁棒性,还给出了特征的有效性检验方法。-Real-time camera system in order to stabilize the output of video, presents a feature-based matching algorithm for

KLT变换对信号的相关性，可以很好去除信号、数据的相关性，减小数据传输量。可以很好的提高速率，在图像处理中也有很好的应用！-KLT transform, signal correlation can be good to remove the signal, the correlation of data, reduce data traffic. May well improve the speed in image processing applications are also very

文章提出了一种惯性辅助的KLT特征跟踪算法，能够提高跟踪的性能-The article proposed a kind of inertia auxiliary KLT feature tracking algorithm can improve the performance of the tracking