Xiaoyang Zhao

I am a research assistant focusing on deep learning–based remote sensing image analysis. My work spans key tasks such as change detection, image classification, and multi-source data fusion, with particular interest in multispectral, hyperspectral, and UAV imagery. I am driven by the challenge of extracting reliable and meaningful information from complex Earth observation data, and by bridging advanced machine learning methods with real-world remote sensing applications. My goal is to develop robust, scalable models that support accurate environmental monitoring and decision-making.

My research focuses on deep learning–based methods for remote sensing image analysis and their applications in extracting information about the Earth's surface. During my master’s studies, I worked on remote sensing image super-resolution, exploring how deep neural networks can enhance spatial details in satellite imagery. My doctoral research focuses on change detection in remote sensing images, aiming to automatically identify patterns of surface change from multi-temporal data. At the same time, I supervise students working on related tasks such as remote sensing image classification. My work involves a range of machine learning approaches, including convolutional neural networks, graph neural networks, few-shot learning, unsupervised learning, and semi-supervised learning. Overall, my research seeks to develop more efficient and robust methods for intelligent interpretation of remote sensing data, supporting large-scale automated analysis and understanding of Earth observation imagery.