A Novel Hyperspectral Image Classification Model Using Bole Convolution with Three-Directions Attention Mechanism: Small sample and Unbalanced Learning

Weiwei Cai, Xin Ning, Guoxiong Zhou, Xiao Bai, Yizhang Jiang, Wei Li, Pengjiang Qian

Abstract: Currently, the use of rich spectral and spatial information of hyperspectral images to classify ground objects is a research hotspot. However, the classification ability of existing models is significantly affected by its high data dimensionality and massive information redundancy. Therefore, we focus on the elimination of redundant information and the mining of promising features and propose a novel bole convolution neural network with a tandem three-directions attention mechanism (BTA-Net) for the classification of hyperspectral image. A new bole convolution is proposed for the first time in this algorithm, whose core idea is to enhance effective features and eliminate redundant features through feature punishment and reward strategies. Considering that traditional attention mechanisms often assign weights in a one-direction manner, leading to a loss of the relationship between the spectra, a novel three-directions (horizontal, vertical, and spatial directions) attention mechanism is proposed, and an addition strategy and a maximization strategy are used to jointly assign weights to improve the context sensitivity of spatial spectral features. In addition, we also designed a tandem three-directions attention mechanism module and combined it with a multi-scale bole convolution output to improve classification accuracy and stability even when training samples are small and unbalanced. We conducted scene classification experiments on four commonly used hyperspectral datasets to demonstrate the superiority of the proposed model. The proposed algorithm achieves competitive performance on small samples and unbalanced data, according to the results of comparison and ablation experiments. The source code for BTA-Net can be found at https://github.com/vivitsai/BTA-Net.