Junde Chen1, Defu Zhang1, Yaser A Nanehkaran1, Dele Li2. 1. School of Informatics, Xiamen University, Xiamen, China. 2. Fujian College of Water Conservancy and Electric Power, Sanming, China.
Abstract
BACKGROUND: As the primary food for nearly half of the world's population, rice is cultivated almost all over the world, especially in Asian countries. However, the farmers and planting experts have been facing many persistent agricultural challenges for centuries, such as different diseases of rice. The severe rice diseases may lead to no harvest of grains; therefore, a fast, automatic, less expensive and accurate method to detect rice diseases is highly desired in the field of agricultural information. RESULTS: In this article, we study the deep learning approach for solving the task since it has shown outstanding performance in image processing and classification problem. Combining the advantages of both, the DenseNet pre-trained on ImageNet and Inception module were selected to be used in the network, and this approach presents a superior performance with respect to other state-of-the-art methods. It achieves an average predicting accuracy of no less than 94.07% in the public dataset. Even when multiple diseases were considered, the average accuracy reaches 98.63% for the class prediction of rice disease images. CONCLUSIONS: The experimental results prove the validity of the proposed approach, and it is accomplished efficiently for rice disease detection.
BACKGROUND: As the primary food for nearly half of the world's population, rice is cultivated almost all over the world, especially in Asian countries. However, the farmers and planting experts have been facing many persistent agricultural challenges for centuries, such as different diseases of rice. The severe rice diseases may lead to no harvest of grains; therefore, a fast, automatic, less expensive and accurate method to detect rice diseases is highly desired in the field of agricultural information. RESULTS: In this article, we study the deep learning approach for solving the task since it has shown outstanding performance in image processing and classification problem. Combining the advantages of both, the DenseNet pre-trained on ImageNet and Inception module were selected to be used in the network, and this approach presents a superior performance with respect to other state-of-the-art methods. It achieves an average predicting accuracy of no less than 94.07% in the public dataset. Even when multiple diseases were considered, the average accuracy reaches 98.63% for the class prediction of rice disease images. CONCLUSIONS: The experimental results prove the validity of the proposed approach, and it is accomplished efficiently for rice disease detection.