基于Harris角点匹配与图像融合算法的嵌入式稳像系统设计  

DESIGNING HARRIS CORNER DETECTION AND IMAGE FUSION ALGORITHM-BASED EMBEDDED IMAGE STABILISATION SYSTEM

作　　者：蔡耀仪[1] 

机构地区：[1]湖南师范大学工学院,湖南长沙410083

出　　处：《计算机应用与软件》2015年第5期220-223,共4页Computer Applications and Software

摘　　要：针对传统电子稳像平台的缺陷,设计一种基于改进Harris角点检测的进行运动估计的FPGA(Field Programmable Gate Array)稳像系统。算法针对FPGA结构特点以及Harris算法理论结构,对其进行必要的改进。针对运动补偿后原有部分图像信息缺失问题,提出一种基于图像融合的缺失图像补偿算法。系统还完成了运动估计和运动补偿部分的硬件设计,使得视频序列帧间存在平移和旋转时能够消除这种运动。研究结果证明,系统能消除随机抖动。For the shortcomings of traditional electronic image stabilisation platform, in the paper we design the FPGA image stabilisation system which estimates the motion based on improved Harris corner detection. Targeted at the structure feature of FPGA and the theoretical structure of Harris algorithm, the new algorithm makes necessary improvements on the system. Taking into account the problem of the infor- mation missing in part of original image after the motion compensation, we propose an image fusion-based missing image compensation algo- rithm. Besides, the system also completes the modules design of the components of motion estimation and motion compensation so that the mo- tion can be eliminated when the translation and rotation are existed in inter-frame of video sequence. Research results prove that the system can eliminate random jitter.

关 键 词：角点检测 现场可编程门阵列 全局运动估算 图像融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]