Rafael Novoa S. A.1 y Guido Herrera M.1
With the objective of evaluating the use of image analysis in the detection of "citrus tristeza virus" on lemon trees (Citrus limon (L.) Burm.) and to calibrate a screening method for the disease, color aerial photographs of sectors with lemon orchards in the Mallarauco Valley; Chile, and samples of healthy plants and plants with virus symptoms were taken. Elisa tests for the virus and analysis of the spectral signatures of the sampled sites were performed. Simple visual inspection of the photographs was enough to discriminate between areas with healthy and sick plants. The reflectivity of the three bands- blue, green and red - was higher in sick plants than in healthy plants. Moreover, the discrimination between sectors of diseased plants was best on the red band. A significant correlation coefficient, 0.63, was found between sectors with ill plants and reflectance for the red band. On correlating the number of diseased plants with the area under curve of the spectral signatures a slightly superior correlation coefficient, 0.64 significant at 95%, was obtained. This implies that the technique not only can be used to discriminate between healthy and ill plants but also to evaluate the percentage of diseased plants in each sector of the orchard.
Key words: remote sensing, image analysis, citrus tristeza virus, lemons trees, diseases.
1 Instituto de Investigaciones Agropecuarias, Centro Regional de Investigación La Platina, Casilla 439 Correo 3, Santiago, Chile. E-mail: email@example.com