Research Interests
Compared with multispectral imaging, hyperspectral imaging is a more advanced technology that can obtain the detailed spectral response of target features.
Real-world use cases that can benefit from such images span multiple fields, including precision agriculture, chemistry, biology, medicine, land cover applications, natural resource management, and natural disaster detection.
Nevertheless, the extraction of meaningful information from hyperspectral remote sensing data is difficult due to its big data characteristics such as large volume, high dimensionality, spatial and spectral heterogeneity and noise effect.
The machine learning and especially the deep learning techniques allow new opportunities for enhanced hyperspectral analytics. The research on hyperspectral remote sensing images processing has entered into a new stage by leveraging these techniques for effective processing and analysis of the acquired hyperspectral data. The current methods mainly focus on different hyperspectral data processing tasks, such as atmospheric correction, reduction of data dimensionality, information extraction, classification and target detection, etc. These methods are still inadequate in terms of speed and accuracy for developing advanced intelligent hyperspectral analytics.
My research focuses on introducing new theories and new methods in deep learning and machine learning to hyperspectral image processing, aiming to comprehensively integrate the advantages of deep learning techniques with the unique characteristics of hyperspectral data to improve both accuracy and speed; in the meantime, my research also highlight key environmental applications of hyperspectral data to address real world applications.
Education
- MSc in Mechatronics Engineering, Shandong University of Technology
- BEng in Mechanical Engineering, Qingdao University