Style Transformation-Based Spatial–Spectral Feature Learning for Unsupervised Change Detection

Due to the inconsistent imaging environment, the styles of multitemporal multispectral images (MSIs) are quite different, such as image brightness and transparency. For multitemporal MSIs with different styles, the “same object with different spectra” problem is one of the biggest challenges in change detection. To overcome the challenge, a novel unsupervised spatial–spectral feature learning (FL) framework based on style transformation (ST) (called STFL-CD) is proposed for MSI change detection in this article. For dual-temporal MSIs, the proposed STFl-CD algorithm consists of two phases: ST and spatial–spectral FL. Since the image styles are inconsistent under different imaging environments, the first innovation is to transform the image styles through unmixing and reconstruction. Through ST, the challenge of the “same object with different spectra” problem will be reduced fundamentally. By introducing the attention mechanism, the other innovation is to extract the joint spectral–spatial change features based on a 3-D convolutional neural network with spatial and channel attention. In addition, for multitemporal MSIs, a multitemporal version STFL-CD (MT-STFL-CD) framework is designed based on a recurrent neural network to learn the correlation features between multitemporal remote sensing images. Both of the visual and quantitative results on the real MSI datasets indicate that the proposed unsupervised STFL-CD frameworks have significant advantages on multitemporal MSI change detection. In particular, the performance of the proposed unsupervised STFL-CD algorithm is even comparable to that of the state-of-the-art supervised or semisupervised methods.

I. Introduction

With the development of remote sensing technology, a great deal of remote sensing data is produced every day. The application of this data has a tremendous impact on social development. However, the intelligence level of interpretation and analysis of remote sensing data is relatively low, which seriously restricts the research and application of remote sensing technology. Therefore, it is very meaningful to explore the advanced intelligent technology on remote sensing image interpretation to solve the “data rich but information poor (DRIP)” problem [1]. As one of the earliest and most widely used fields in remote sensing image processing, change detection has wide applications, such as agricultural production evaluation, urban change monitoring, disaster rescue, and military reconnaissance [2]. For example, when large-scale natural disasters occur, such as earthquakes and floods, fast and precise change detection analysis of remote sensing data can guide the allocation of rescue resources and disaster assessment. Therefore, remote sensing image change detection is an extremely important research topic, which attracts the wide attention of researchers [3].