摘要
提出了一种改进特征点匹配算法的电子稳像技术.通过合理的区域划分,进行图像的特征点提取,并采用距离不变准则,对特征点的有效性进行验证,利用卡尔曼滤波的方法来进行运动补偿.实验结果表明:此算法具有良好的稳像效果.
Presents an improved features tracking algorithm in electronic image stabilization. Through dividing image into many blocks in reason, the suitable feature points can be selected, propose the distance invariant criterion to check the feature matching, Kalman filter is applied to motion compensation to achieve the stability of the image sequence. The results of experiment show that this method has good performances in image stabilization.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第3期126-129,共4页
Microelectronics & Computer
基金
江苏省高校自然科学基础研究项目(06KJA14003)
关键词
电子稳像
特征提取
区域分割
特征匹配
运动估计
electronic image stabilization
feature extraction
region segmentation
feature matching
motion estimation
