Rescaling-Assisted Super-Resolution for Medium-Low Resolution Remote Sensing Ship Detection

Medium-low resolution (M-LR) remote sensing ship detection is a challenging problem due to the small target sizes and insufficient appearance information. Although image super (SR) has become popular solution in recent years, ability of SR limited since much information lost input images. Inspired by powerful embedding encoder rescaling, this paper, we introduce rescaling guide training SR. Specifically, add an adaption module before network, use pre-trained network optimization module. In way, more embedded adapted M-LR images, subsequent can utilize achieve better performance. Extensive experimental results demonstrate effectiveness our method on More importantly, be used as pre-processing approach improve