Spectral Characteristics of Penicillium Species using a Frequency controlled Liquid Crystal Filter

• Published in: Biosystems engineering 2004-07, Vol.88 (3), p.265-269
• Main Authors: Okushima, L, Saito, M, Sase, S, Ling, P, Ishii, M
• Format: Article
• Language: English
• Subjects: Agricultural machinery and engineering; Agronomy. Soil science and plant productions; Biological and medical sciences; Euclidian distances; Fundamental and applied biological sciences. Psychology; Generalities. Biometrics, experimentation. Remote sensing; multispectral data; Penicillium; Solanum tuberosum; spectral analysis; taxonomy
• ISSN: 1537-5110; 1537-5129
• DOI: 10.1016/j.biosystemseng.2004.04.001

The use of multi-spectral data in the identification of Penicillium species was investigated, using 19 isolates of 13 species. For one isolate, 34 narrow band images were captured between 420 to 750 nm, 10 nm apart after the 1-week cultivation of the malt extract agar (MEA) and potato dextrose agar (PDA) media. The Euclidean distance of mean grey level intensities on the back surfaces presented better similarity in the same species than that on the front surfaces presented. The Euclidean distance made it possible to reduce the 18 candidates in the same species to 1–4 nominations. As an additional spectral character, Euclidean distances of the pixel areas over 1600 and 2300 grey level intensities on the image of 750 nm were estimated. The Euclidean distance of pixel areas in the same species showed least value in the nominations. The spectral characteristics were useful in reducing the number of candidates for the prerequisite before the identification.