一种聚类与灰度投影相结合的电子稳像方法  

An Electronic Image Stabilization Algorithm Based on Clustering Combined with Gray Projection

作　　者：赵猛[1] 曹茂永[1] 滕升华[1] 

机构地区：[1]山东科技大学信息与电气工程学院,山东青岛266590

出　　处：《山东科技大学学报（自然科学版）》2013年第2期85-89,共5页Journal of Shandong University of Science and Technology(Natural Science)

基　　金：山东省自然科学基金项目(ZR2010FM027);山东科技大学科研创新团队支持计划项目(2010KYTD101)

摘　　要：提出了一种聚类与灰度投影技术相结合的电子稳像方法。首先,利用小波变换法提取特征点,并计算其方向角,通过聚类方法估计旋转角度;角度参数补偿后,根据图像的灰度信息特征得到投影曲线,通过投影曲线间的相关计算进行平移量的运动估计。实验结果表明,该方法对于摄像机的不规则运动参数估计是有效的,而且能够做到实时处理。This paper presented an electronic image stabilization method of combining clustering with gray projection algorithm. At first, the wavelet analysis technique was used to extract feature points and calculate the direction angle, and the rotation angle was estimated by a clustering method. After angle parameters were compensated, the projection curve was obtained according to gray information characteristics of the image, and then the motion estimation of the shift parameters was carried out by gray projection algorithm. The experimental results show that this method is effective for estimating parameters of irregular movements of the camera and can also work in real-time processing.

关 键 词：特征提取 运动估计 灰度投影 电子稳像 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]