Single Space Object Image Super Resolution Reconstructing Using Convolutional Networks in Wavelet Transform Domain

With the increasing importance of space exploration, the research of space object is becoming more and more important because high-quality space object images are meaning for space attack and defense confrontation. However, high-quality space object images are very difficult to obtain because of the large number of various rays in the space environment and the inadequacy of optical lenses and detectors on satellites to support high-resolution imaging. Image super resolution reconstruction methods are the most cost-effective way to solve the problem. In this paper, we propose a deep convolutional neural network based method to improve the resolution of space object image. The implementation of our method is in wavelet transform domain rather than spatial domain because wavelet transformation could decompose different frequencies of the image very effectively and this could further more enhance the performance. The experiment result shows that our method could achieve a very good performance.