AUTHOR=Feng Lei , Wu Baohua , He Yong , Zhang Chu 

TITLE=Hyperspectral Imaging Combined With Deep Transfer Learning for Rice Disease Detection

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2021.693521

DOI=10.3389/fpls.2021.693521

ISSN=1664-462X

ABSTRACT=Various rice diseases threaten the growth of rice. It is of great importance to achieve rapid and accurate detection of rice diseases for precise disease prevention and control. Hyperspectral imaging was used to detect rice leaf diseases with four different varieties of rice. Considering that it costs much time and energy to develop a classifier for each variety of rice, deep transfer learning was first developed for rice disease detection across different rice varieties. Three deep transfer learning methods were adapted for twelve transfer tasks, including fine-tuning, Deep CORrelation ALignment (Deep CORAL), and Deep Domain Confusion (DDC). A self-designed CNN was set as the basic network of deep transfer learning methods. Fine-tuning achieved the best transferable performance with an accuracy of over 88% for the test set of the target domain in the majority of transfer tasks. Deep CORAL obtained an accuracy of over 80% on four of all transfer tasks, which was superior to DDC. A multi-task transfer strategy has been explored with good results, indicating the potential of both pair-wise and multi-task transfer. The saliency map was used for visualization of key wavelength range captured by CNN with and without transfer learning. The results indicated the wavelength range with and without transfer learning was overlapped to some extent. Overall, the results suggested deep transfer learning methods could achieve rice disease detection across different rice varieties. Hyperspectral imaging in combination with the deep transfer learning method was a promising solution for efficient and cost-saving field detection of rice diseases among different rice varieties.