Abstract: For aerial image sequence, a new method is proposed to estimate the motion vectors of the image feature points, which is extracted by means of improved Minimum Intensity Change(MIC) algorithm. Adaptive Rood Pattern Search(ARPS) algorithm is applied for finding the matching points of the feature points. Combining Random Sample Consensus algorithm(RANSAC), the affine transform parameters of the background motion can be estimated, and then the background motion can be compensated. The improved background motion compensation method offers great promise in the moving target detection of aerial image sequence.

Key words: background motion compensation, aerial image, moving target detection, Minimum Intensity Change(MIC)

摘要： 针对航拍图像序列提出用于提取图像特征点的最小亮度变化算法。利用特征点进行运动矢量估计，使用自适应十字模板搜索算法找到各个特征点的匹配点，结合随机样本一致算法估计背景运动模型的仿射变换参数，实现一种改进的背景运动补偿技术。该技术保障了航拍图像中运动目标检测的有效性。

关键词: 背景运动补偿, 航拍图像, 运动目标检测, 最小亮度变化

CLC Number: 

• TN911.73