中文摘要：

针对弹载合成孔径雷达（SAR）距离徙动补偿困难以及原始后向投影（BP）算法计算量大实时性差的问题，本文提出了一种基于解线频调（dechirp）弹载SAR的改进BP成像算法．首先建立导弹在末制导阶段的dechirp回波信号模型并完成距离向的聚焦，然后对距离向进行条带划分，结合子孔径合并和图像分裂技术将每个条带内的回波反向投影至成像区域进行相干累加，从而得到成像区域的SAR图像，最后分别采用原始BP算法与本文改进BP算法进行模拟和实测对比实验．实验结果表明，该算法能够对目标区域精确成像；由于各条带间的成像是相互独立的，易于算法的简化及并行处理；极大地降低了算法的运算量、提高了运算速度，增强了算法的工程可实现性．该研究成果在遥感，目标的探测与识别，精确制导等领域中具有重要的应用价值．

英文摘要：

The range migration of missile borne synthetic apertare radar （SAR） is very difficult to correct using all the conventional imaging algorithms but the back projection algorithm; however, the large computation burden and the low efficiency are the key problems existing in its practical application. To solve the problems, an extended back projection imaging algorithm for the dechirped missile borne SAR is proposed. Firstly, a new signal model is built for the dechirped missile borne SAR, and the range compression is implemented in the range frequency domain. Secondly, the imaging region is divided into several stripes in range direction, and the raw echo data is back projected to each stripe for coherent integration by using the sub-aperture merging and image splitting technique. Finally, the entire SAR image can be obtained by combining the subimages of all the stripes. The validity of the proposed algorithm is demonstrated by the simulated and real SAR datasets. Testing results indicate that the extended algorithm is appropriate for achieving accurate dechirped missile borne SAR image. Moreover, it can be easily parallelized because the stripe imaging is independent of each other, so that can greatly decrease the cofnputation burden, and improve the computation speed. The method introduced in this paper has important theoretical significance in realistic remote sensing, detection and recognition of military targets and precision guide.