相邻TDI CCD卫星影像的匹配拼接

Stitching Algorithm for Sub-Images of Adjacent TDI CCD Chips Based on Image Matching

为了获取较大的卫星影像幅宽,通常在传感器焦平面内交错排列多个时间延迟积分(TDI)CCD。由于分片影像具有同源性、同时性和一定的重叠度,采用图像匹配的方法选取大量连接点可以较好地将影像拼接起来。首先分析了TDI CCD成像的主要影响因素;然后在标准影像上提取出大量的、均匀分布的特征点,利用成像设计参数预测同名点位置;最后采用垂直轨道方向直接平移、沿轨道方向分段改正的重采样策略获得拼接影像,有效地避免了因地形起伏带来的拼接错位。实验结果表明,采用该方法处理后的分片影像拼接效果良好,满足后续处理的要求。

For reaching a sufficient swath width, some TDI CCD chips are combined in the focal plane. The images generated by the individual TDI CCD are simuhaneous, geographical homologous and overlap slightly and so they can be stitched by tie points which are selected based on image matching. In this paper, some key factors that af- fect the imaging quality of TDI CCD images are discussed firstly. Then, a large amount of uniform distributed fea- ture points are extracted in the standard image and the homonymy points can be calculated by the design parame- ters together with proper image matching algorithms. Finally, the image before stitching are shift transformed in the cross-track direction and resampled with different parts in the along-track direction to avoid mistakes which caused by topographic relief. Experiment results show that no position mistakes can be found in the stitched image pro- cessed by this method and it meets the requirements of the subsequent processing.