作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2013, Vol. 39 ›› Issue (3): 300-305,310. doi: 10.3969/j.issn.1000-3428.2013.03.060

• 开发研究与设计技术 • 上一篇    下一篇

基于Harris角点与改进Hu矩的电子稳像算法

吴国楠1,周超超2,尹文波1   

  1. (1. 复旦大学专用集成电路与系统国家重点实验室,上海 201203;2. 上海汽车信息产业投资有限公司,上海 200041)
  • 收稿日期:2012-05-03 出版日期:2013-03-15 发布日期:2013-03-13
  • 作者简介:吴国楠(1986-),男,硕士研究生,主研方向:数字图像处理,集成电路设计;周超超,工程师;尹文波,高级工程师、硕士

Electronic Image Stabilization Algorithm Based on Harris Corner and Modified Hu-moments

WU Guo-nan  1, ZHOU Chao-chao  2, YIN Wen-bo  1   

  1. (1. State Key Laboratory of ASIC & System, Fudan University, Shanghai 201203, China; 2. Shanghai Automotive Information Systems Co., Ltd., Shanghai 200041, China)
  • Received:2012-05-03 Online:2013-03-15 Published:2013-03-13

摘要: 为提高抖动视频电子稳像系统的稳像效果,提出一种基于Harris角点和改进Hu几何不变矩的电子稳像算法。该算法检测视频每帧图像的Harris角点为特征点,并计算其邻域图像改进的Hu矩作为对应特征向量,以仿射变换为模型,通过特征点匹配计算前后帧的相对运动参数,从而计算出图像序列的运动轨迹,加以相应的运动平滑补偿,得到保留镜头正常运动信息稳定的视频图像序列。实验结果表明,在视频图像帧间存在旋转、平移和轻微缩放的情况下,该算法能有效降低特征匹配的计算量,且保持良好的匹配效果。

关键词: 电子稳像, Hu矩, Harris角点, 运动估计, 仿射变换, RANSAC算法

Abstract: To improve the efficiency of the video stabilization system, an algorithm based on Harris and modified Hu-moments is proposed. In this algorithm, Harris corners of each frame are detected as the feature-points, and the modified Hu-moments in the corner-centered neighbourhood are calculated to be the feature-vectors corresponding to the feature-points. The affine transformation model is used in the motion estimation between the frames. The motion traces of each direction are got. A stabilized video sequence is generated after compensation being exerted on the frames. Experimental results show that when the images are rotated or slightly zoomed, and the compute complexity is greatly reduced in the processing of feature-matching with an ideal result.

Key words: electronic image stabilization, Hu-moments, Harris corner, motion estimation, affine transformation, RANSAC algorithm

中图分类号: