Quantification and localisation of individual leaf disease lesion for grading severity of late blight

Abstract

Detecting incidence and grading the severity of plant diseases caused by pathogens is among the essential activities in precision agriculture. This research proposes novel noetic integration between pathology and advanced yet straightforward image processing technique for grading the severity of vegetable late blight. Until recently, most of the presented image processing techniques had been, and some still are, grading severity based on the visual understanding of disease symptom as a single lesion colony. One of the most significant advantages of the proposed method is quantifying and localising disease symptom colonies into symptomatic and necrotic in accordance with pathological disease analogy for actual severity grading. In the 1st phase of the study, individual symptomatic (RS), necrotic (RN), and blurred (RB, in- between healthy and symptomatic) regions were identified and segmented. The isolated diseased lesions are then quantified and localised for correlationwith a standard area diagram which gives the accurate grading of disease severity. Results obtained indicated great potential for accurate grading by which pathological knowledge and advance computer network operate in proper synergy. It is also envisaged that this research method would provide meaningful insight into the critical current and future role pathological integration in machine learning for food security.