![]() proposed the regularized deep Fisher mapping (RDFM) method per Fisher criterion to enhance the feature separability by using the neural network algorithm, which can eliminate the overfitting problem. proposed an unsupervised greedy method based on Deep Belief Network (DBN), which trains layer by layer, solving the vanishing gradient problem caused by deep training. In recent years, the theory of deep learning has set off a wave in the field of pattern recognition. Such results are promising in comparison with other state-of-the-art methods. The experiments are conducted on the Airborne Synthetic Aperture Radar (AIRSAR) data of Flevoland and the results show that the classification accuracy on 14 types of land cover is up to 98.56%. ![]() Then, the Softmax classifier is employed to classify these features. These extracted features are first combined into a fully connected layer sharing the polarization and spatial property. The other is utilized to extract the spatial features of a Pauli RGB (Red Green Blue) image. The proposed method is built on two deep CNNs: one is used to extract the polarization features from the 6-channel real matrix (6Ch) which is derived from the complex coherency matrix. In this paper, a novel method based on a dual-branch deep convolution neural network (Dual-CNN) is proposed to realize the classification of PolSAR images. Existing polarimetric synthetic aperture radar (PolSAR) image classification methods based on the CNN only consider the polarization information of the image, instead of incorporating the image’s spatial information. The deep convolution neural network (CNN), which has prominent advantages in feature learning, can learn and extract features from data automatically.
0 Comments
Leave a Reply. |