大序列无人机影像畸变差快速修正方法

Fast Correction Method for Image Distortion of Large Sequence UAV

分析了数字影像畸变差修正模型与GPU并行计算技术,提出一种大序列无人机影像畸变差快速修正方法。通过GPU线程格网选择配置与多层次性存储等优化措施,本方法较传统CPU串行算法效率提升了6.25倍,节约了大序列无人机影像的畸变差修正的时间。

In order to correct the distortion in a large sequence of UAV images efficiently,we studied the distortion correction model of the digital image and the GPU parallel computing technology,and proposed a fast correction method for image distortion of large sequence UAV.Selecting the best grid configuration and multi-level memory optimization were used in this method to accelerate the process of images’ aberration correction.Experimental results show that compared with the traditional CPU serial algorithm,the efficiency of this method is improved by 6.25 times,and the distortion correction time of large serial UAV images is saved.