Image - Based Plant Disease Detection

Abstract

Agriculture productivity is one of the major sectors of the Vietnam economy, which contributes roughly 20 percent to the GDP and food to fulfill 95.54 million Vietnamese people [1]. For thousands of years, the Vietnamese are practicing agriculture. However, some small farmers use the outdated traditional agriculture method, which is highly sensitive to climate change and weather extremes such as the increase of severe drought episodes and salinity intrusion, which will pose challenges for farmers. Nowadays, the number of high-quality laborers in the agriculture, forestry, fishery sector is decreasing, and the negative effect of climate change on agriculture creates many kinds of crop disease, which constitute a significant threat to crop plan on the bigger picture causes serious implication for Vietnam economic development. Applying the new computer technology and advanced equipment to develop the automatic leaf disease detection technique helps and provides useful information for the farmer to identify diseases. This thesis presents image processing techniques such as image segmentation, image classification combined with the smartphone's advanced HD camera to propose a new way for smartphone-assisted plant disease determination.