Cognitive Fusion of Graph Neural Network and Convolutional Neural Network for Enhanced Hyperspectral Target Detection

In recent years, deep learning has emerged as a prominent technique in hyperspectral target detection (HTD). Extensive research has highlighted the potential of graph neural network (GNN) as a promising framework for exploring non-Euclidean dependencies within hyperspectral imagery (HSI). However, GNN has not been introduced to HTD. Additionally, achieving a balanced training set while effectively suppressing background remains a challenge. Therefore, we propose the cognitive fusion of GNN and convolutional neural network (CNN) for enhanced HTD (named as CFGC), which marks the first integration of GNN and CNN in HTD. Initially, using sparse subspace clustering (SSC) and a similarity measurement strategy, we select the most representative background samples for HTD. Subsequently, linear interpolation combines the prior target with the Laplacian-weighted prior target, yielding abundant targets with meaningful transformations. Finally, a fused network of CNN and GNN is utilized for training both the prior target and the constructed training set. Significantly, the incorporation of attention mechanism in both the CNN and GNN branches stands out as a noteworthy advantage, augmenting the models’ ability to selectively prioritize crucial information. Four benchmark hyperspectral images have been used in extensive experiments, and the results demonstrate that CFGC exhibits superior performance in HTD.

I. Introduction

Hyperspectral imagery (HSI) with high spatial and spectral resolution is a 3-D image that simultaneously acquires spectral information in multiple narrow bands [1], [2], [3]. Therefore, HSI is widely used for tasks like ground object classification [4], [5], [6], change detection [7], [8], anomaly detection [9], [10], [11], [12] target detection [13], [14], [15] spectral unmixing [16], [17], and so on. Hyperspectral target detection (HTD), which identifies specific objects by extracting unique spectral features from given data with a prior target, is currently receiving increasing attention [18], [19]. Due to its ability to provide fine discrimination and detailed analysis, the task finds applications in various fields, including environmental monitoring [20], agriculture [21], geological exploration [22], military reconnaissance [23], and so on.