Detection and Classification of Mosaic Virus Disease in Cassava Plants by Proximal Sensing of Photochemical Reflectance Index

Cassava Mosaic virus Disease (CMD) is the most severe and widespread virus infection that affects cassava (Manihot esculenta Crantz) crops. This paper investigates the application of photochemical reflectance index (PRI) imaging to detect and assess the impact of varying levels of CMD infection in cassava. Towards this, narrow band reflectance images of field-grown cassava plants were recorded at 531 and 571 nm by proximal sensing with a multispectral imaging system (MSIS). It was observed that the PRI values increase with increasing levels of CMD infection in all the varieties of cassava studied. A scatter plot of the PRI image intensity yielded a sensitivity of 85 % and specificity of 79 % for discriminating visibly no CMD from initial CMD and a sensitivity of 93 % and specificity of 92 % for discriminating initial CMD from advanced CMD. Area under the receiver operator characteristics (AUC-ROC) curve was used to discriminate the CMD infection level by differentiating visibly no CMD from initial CMD [AUC = 0.92] and initial CMD from advanced CMD [AUC = 0.99]. It was observed that PRI values determined from the experimental data follow a linear inverse relationship with net photosynthetic rate (Pn) (R 2 = 0.76) and total leaf chlorophyll (Chl) content (R 2 = 0.80). The results show that PRI imaging can be utilized to discriminate healthy plants from CMD and other stress infected crops by proximal sensing in outdoor plants.