The geometric rectification of linear pushbroom imagery is the performance bottleneck for data processing of massive remote sensing image . The back projection is essential for photogrammetric processing of linear pushbroom imagery such as epipolar resampling and DOM generation and its main objective is to search the best scan line .The traditional best scan line searching strategy based on iterative calculation of collinearity equation in image space needs a lot of processing time which has little practical use .A fast algorithm for best scan line searching based on the geometric constraint of the central perspective plane of scan line is proposed in this paper . The best scan line is determined efficiently within several iterations in ground space by calculating the distance between ground point and the central perspective plane in object space .Due to the installation error and lens distortion ,the CCD pixels of linear array scanners are not in a straight line .Accordingly , the CCD array is partitioned into several line segments and the determination of central perspective plane of the line segment correspondenttothegroundpointisdiscussed.TheADS40airbornelinearpushbroomimageryandtheMarsExpress HRSC satellite imagery are tested .Compared with the traditional method in image space ,the method proposed in this paper shows better results in best scan line searching and differential rectification .%线阵推扫式影像的几何纠正是当前海量遥感影像数据处理的难点与性能瓶颈。基于严密几何模型对线阵影像进行微分纠正是生成各级影像产品的基础，其核心是地面点的反投影，即确定地面点对应的最佳扫描行。传统的最佳扫描行搜索算法基于像方的共线方程迭代计算，效率低，实用化程度不高。本文基于严密几何模型提出一种适用于大数据量线阵影像微分纠正的最佳扫描行快速搜索策略，利用扫描行投影面进行物方几何约束，通过判断物方点到扫描行投影面的距离定位最佳扫描行。由于线阵列CCD像元并非严格按照一条直线排列，笔者讨论了CCD线阵的分段处理以及分段投影面的确定方法。利用机载ADS40、火星快车HRSC线阵影像进行试验。结果表明，本文方法仅需少量几次物方空间解析几何迭代计算即可精确定位最佳扫描行，计算效率相比基于像方的搜索方法提高了8倍以上，线阵影像微分纠正速度提升了6倍以上。
展开▼
