A Enhanced DeepLabv3+ for PolSAR image classification

Polarimetric synthetic aperture radar (PolSAR) is an advanced remote sensing technology that can provide radar images with polarimetric information. The research of PolSAR image classification algorithm has always been one of the hot spots and difficulties in SAR remote sensing image processing. Deep learning algorithms usually require a large amount of training data to achieve a good classification effect. However, in the field of PolSAR image classification, the available data set is relatively small due to the complexity of data acquisition and processing. In order to improve the effectiveness of deep learning algorithm in PolSAR image classification. In this chapter, the idea of SKNet unit is integrated into DeepLabV3+network, and a SK_ ASPP unit is proposed, which can adaptively learn feature weights, so as to better extract features related to classification tasks. Experimental results on Flevoland data sets show that the network has better accuracy, overall accuracy (OA) and Kappa accuracy than the original DeepLabv3+.