Submitted to: Phytopathology
Publication Type: Abstract Only
Publication Acceptance Date: April 1, 2005
Publication Date: June 1, 2005
Citation: Bock, C.H., Parker, P.E., Cook, A.Z., Gottwald, T.R. 2005. Image analysis versus visual assessment of incidence and severity of citrus canker symptoms. Phytopathology. 95:6S11. Technical Abstract: <i>Xanthomonas axonopodis</i> pv <i>citri</i> causes citrus canker. Disease assessment is important for monitoring epidemics. Visual assessment (VA) is presently the only reliable means of detection. To investigate how VA of symptoms compared to image analysis we used digital images of 214 citrus leaves with a range of incidence and severity of citrus canker. Comparison of assessments by three plant pathologists (VA1-3) versus image analysis (IA) showed a poor relationship between lesion number and % area infected (R<sup>2</sup>=0.34-0.60). GLM analysis demonstrated no significant differences between IA and the visual assessors for lesion number (mean 16.3-18.7 lesions per leaf), but there were significant differences in the severity score for % area necrotic (3.3-5.4% mean area) but not for the total area infected (chlorotic+necrotic, 8.9-11.0% mean area), suggesting VA tended to overestimate the necrotic area, but was more accurate when assessing the whole infected area. The variability in VA for all raters was similar across all range categories of disease severity and lesion number. A sample of 53 leaves was converted to black/white images and IA/VA performed. The results were similar to using color images and both IA and VA provided comparable methods of assessment. IA appeared to provide an objective approach to assessing canker infected leaves.