基于深度学习的多帧遥感降质图像三维重建算法OACSTPCD

In order to improve the contrast and image quality of multiframe degraded remote sensing images,a deep learning based 3D reconstruction algorithm for multiframe degraded remote sensing images is proposed.The trigonometric function transformation method combined with high pass filter is used to enhance the contrast of multiframe degraded remote sensing images.Based on a generative adversarial network that includes generators and discriminators,a self attention layer is introduced into the discriminator,and a residual module of the self attention mechanism is designed to generate a self attention generative adversarial network model.The enhanced image input model is learned and trained to obtain global features of multiple degraded remote sensing images,and then three-dimensional reconstruction of multiple degraded remote sensing images is achieved.The testing results show that the algorithm has good ability to enhance multi frame degraded remote sensing images and improve image contrast.The permeability indexs(PI)are all above 0.92,and the reconstruction effect is good.