Application of infrared thermal imaging technique and discriminant analysis methods for non-destructive identification of fungal infection of potato tubers

Abstract

A reliable, nondestructive and fast method based on active thermography has been evaluated in the present research to detect Fusarium solani fungus in potato tubers. This fungus causes the main disease in potato as dry rot. Healthy potato tubers and fungal contaminated ones at different stages (1 to 7 day after infection) have been provided to be classified. Different heating/cooling treatments including two heating temperature levels (60 and 90 °C) and four cooling time levels (10, 25, 40 and 70 s) were considered to find optimum thermography condition. The thermal images were acquired using a thermography camera and a heating box. The treatments were analyzed in a factorial experiment based on completely randomized design and heating temperature of 90 °C and cooling time of 40 s were selected as optimum thermography condition. Some statistical features were extracted and classified by linear and quadratic discriminant analysis methods. The classification results demonstrated that quadratic discriminant analysis can classify the healthy and contaminated tubers at different stages with high accuracy (98.5%).