论文部分内容阅读
光束法平差是遥感图像实时拼接系统的关键步骤,它需要消耗大量的运算时间,同时也决定了系统的拼接精度。针对画幅扫描红外成像实时拼接系统中的光束法平差算法,提出了一种实用的光束法平差策略,利用相邻扫描行图像的配准平差,将最新扫描行图像依次配准到已拼接图像中,为拼接图像的实时更新奠定了基础。同时利用GPGPU对光束法平差算法进行了并行加速,在保证精度的情况下,提高了拼接算法速度。通过画幅扫描成像的实时平差实验,验证了算法的精度和实时性。
The beam method adjustment is the key step in the real-time splicing system of remote sensing images. It needs to consume a large amount of computing time and also determines the splicing accuracy of the system. Aiming at the beam method adjustment algorithm in frame scanning infrared imaging real-time splicing system, a practical beam method adjustment method is proposed. By using the registration adjustment of adjacent scanning line images, the latest scanning line images are sequentially registered to Stitching images, laid the foundation for real-time updates splicing images. At the same time, the GPGPU is used to accelerate the beamforming algorithm in parallel, which improves the speed of the stitching algorithm under the condition of ensuring accuracy. The real-time adjustment experiment of frame scan imaging verifies the accuracy and real-time performance of the algorithm.