Cassava leaf NDVI - Artificial neural network (CaNDVI-ANN): A low cost, portable and non-invasive cassava plant health monitoring device

Abstract

Numerous types of plant diseases could significantly impact the gross yield of food and agricultural assets wherein it could possibly result in a huge number of losses in terms of quantity and quality as well as in the economic aspect. Known practices to minimize loss because of plant infections includes identifying initial symptoms of infections through either based on visual examination or through laboratory investigations. However, these practices require a large workforce and take too much time. In order to solve the deficiencies of these usual and traditional approaches, this study presents a low cost, a portable and, noninvasive plant health monitoring device using an IR camera and an artificial neural network for image classification. The initial prototype was tested on a cassava plant to assess its photosynthetic activity which greatly indicates the current health status of the plant. The test result shows that almost all of the positive test samples were correctly identified, meaning it has identified low false-negative results, which obtained an accuracy of 89 %.