Title:Towards automated mobile-phone-based plant pathology management

Abstract:This paper presents a framework which uses computer vision algorithms to standardise images and analyse them for identifying crop diseases automatically. The tools are created to bridge the information gap between farmers, advisory call centres and agricultural experts using the images of diseased/infected crop captured by mobile-phones. These images are generally sensitive to a number of factors including camera type and lighting. We therefore propose a technique for standardising the colour of plant images within the context of the advisory system. Subsequently, to aid the advisory process, the disease recognition process is automated using image processing in conjunction with machine learning techniques. We describe our proposed leaf extraction, affected area segmentation and disease classification techniques. The proposed disease recognition system is tested using six mango diseases and the results show over 80% accuracy. The final output of our system is a list of possible diseases with relevant management advice.