Automated Sugarcane Disease Detection Using Faster RCNN with an Android Application
R. Murugeswari, Z. Sharik Anwar, V. Raja Dhananjeyan, C. Naveen Karthik
Abstract
Sugarcane is a crop that is grown vastly in our country. Diseases are common in plants and so in sugarcane as well. Visual identification of sugarcane diseases is the method commonly used by a farmer to detect and identify them. This technique takes more time and becomes difficult in this process in huge farms. If not treated earlier, it causes a threat both to the farmers and the sugarcane industry. Detecting them in early stages using Deep Convolutional Neural Network can prevent losses. This work attempts to combine Convolutional Neural Network architectures of Faster Region-based Convolutional Neural Network to improve detection and recognition of sugarcane illnesses.1500 pictures of healthy sugarcane leaves and afflicted sugarcane illnesses were used to train the models. An android application also serves as the end user interface for capturing photographs and quickly determining the disease's and it's accuracy.