机载电子稳像技术研究，很好的一篇博士论文。

本章首先分析了航摄动态图像的特点，并在对机载航摄图像序列进行大量的实

验和理论分析的基础上，提出了适用机载视频图像序列稳定的有效算法，即灰度投影

算法和基于特征点匹配算法。针对灰度投影算法易受图像的背景信息影响的特点，

详细地讨论了如何提高该算法用于机载电子稳像的精度，并通过实验验证了所提方

法的有效性。在长焦距成像系统中，旋转运动对图像序列的稳定有着重要的影响，

机载成像系统在成像过程中，不仅有平移运动，还包含旋转运动，必须考虑对旋转

运动的补偿，提出了用特征点匹配算法应用于机载电子稳像系统图像序列帧间旋转

运动和平移运动的检测，为了保证该算法的有效性，先提取图像的边缘特征，再采

用分区的方法，在各个分区内按一定的扫描顺序选取灰度最大值点作为特征点，用

于全局运动检测。帧间运动的补偿方法也是影响图像序列平稳性的一个重要因素，

本文提出用均值滤波的方法来确定各帧的补偿量，可避免图像序列的跳跃，增强动

态图像序列的平稳性。

-Airborne electronic image stabilization technology research, a good one doctoral dissertation.

This chapter first analyzes the characteristics of aerial dynamic image, and in the airborne aerial image sequence a lot of real-

Experience and theoretical analysis based on the proposed application of airborne video image sequence of a stable and effective algorithm, namely, gray projection

Algorithm and matching algorithm based on feature points. Projection algorithm for gray-scale image of the background information vulnerable to the impact of the characteristics of

Discussed in detail how to improve the algorithm for on-board electronic image stabilization, precision, and through experimental verification of the proposed side

Effectiveness of the method. In the long focal length imaging system, the rotation on the stability of image sequences has important implications

Airborne Imaging System in the imaging process, not only translational motion, but also includes rotary move