Performance Analysis Of Hyperspectral Data Classification Based On Hybrid Neural Network Approach

Abstract

Hyperspectral Image (HSI) data classification is a challenging task in remote sensing data analysis, which has been applied in many domains for better identification and inspection of the earth surface by extracting spectral and spatial information. Recent advances in neural networks have made great progress in the HSI classification. However, many traditional methods are based on handcrafted features, which brings difficulties for multi-classification tasks due to spectral intra-class heterogeneity, similarity of inter-class and higher model complexity. Consequently, conventional classifiers are not feasible to extract distinctive features. In order to improve the classification, this paper presented a hybrid neural network approach with pretrained DNN model with ANN based method. The performance of proposed classifiers is compared and found to be effective, Overall improvement in a classification performance is 4.6% in comparison with existing methods. Though, this model could be applied and validated on geological mapping and urban investigation in terms of live hyperspectral image dataset.