【摘要】 遥感成像的超分辨重建对于提高遥感对象识别成功率非常重要,但在图像重建中会出现样本训练需求量大,从而影响算法效率及重建分辨率等问题。对此,提出单一遥感图像的高低分辨影像字典学习超分辨重建算法。首先,基于现有高分辨遥感影像,经过预处理操作获得具有高低分辨特征的样本训练集;其次,设计稀疏字典联合学习训练方式,并对具有高低分辨特征的样本训练集进行处理,进而获得高低分辨稀疏字典;最后,基于荻取的高低分辨稀疏字典对高分辨遥感影像进行重建,并对所设计算法进行复杂度分析。仿真对比显示,该算法对于字典训练样本需求量较少,并可获得较好的重建效果。更多还原

【Abstract】 Super resolution remote sensing image is important for the improvement of the target recognition rate,but there is a large demand of training samples in the super resolution reconstruction process,which results in the problem of low computing efficiency and poor reconstruction resolution,so here a super resolution sparse reconstruction algorithm with the high and low resolution dictionary for single remote sensing image is proposed.Firstly,the training set of samples with high and low resolution is obtained from the high resolution remote sensing image based on the existing high resolution remote sensing image;Secondly,here the sparse dictionary joint training method is designed,and the high and low resolution feature of the sample training set is dealt with,which could get the dictionary with high and low resolution;Finally,the high resolution remote sensing image is reconstructed based on the acquired high and low resolution dictionary,and then the complexity of the algorithm is analyzed.The simulation results show that the algorithm has less requirements on the dictionary training sample,and can achieve a good result.更多还原