Deep-Clustering Based Plant Disease Segmentation Network

Abstract

Plant disease is a major factor that reduces the yield of plant cultivation. To solve this problem, many CNN-based disease detection models have been studied. However, existing methods focus on detecting disease regions of plants with a clean or constant background of image, so they are not practical in actual fields. Field images captured with UAVs frequently suffer from complex backgrounds. To overcome this problem, we propose a CNN-based plant disease segmentation network based on deep clustering.