Click on “Download PDF” for the PDF version or on the title for the HTML version.
If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.
Hyperspectral Spectroscopy for Detection of Early Blight (Alternaria solani) Disease in Potato (Solanum tuberosum) Plants at Two Different Growth Stages
Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.orgCitation: 2015 ASABE Annual International Meeting 152186976.(doi:10.13031/aim.20152186976)
Authors: Daniel Atherton, Dennis G. Watson, M. Zhang, Z. Qin, X. Liu
Keywords: Remote Sensing, Early Blight, Spectroradiometer, Potato
Abstract. Early detection of disease and insect infestation within crops and precise application of pesticides can help reduce potential production losses, reduce environmental risk, and reduce the cost of farming. The objective of this study was the detection of early blight (Alternaria solani) in potato (Solanum tuberosum) plants at two different growth stages using a handheld hyperspectral spectroradiometer. Hyperspectral reflectance spectra were captured 10 times over five weeks from plants grown to the vegetative and tuber bulking growth stages. The spectra were analyzed using principal component analysis (PCA), spectral change (ratio) analysis, and partial least squares (PLS) analysis. PCA successfully distinguished more heavily diseased plants from healthy and minimally diseased plants using two principal components. Spectral change (ratio) analysis provided wavelengths (505, 510, 640, 665, 690, 750, and 935 nm) most sensitive to early blight infection followed by ANOVA results indicated a highly significant difference (p < 0.0001) between disease rating group means. In the majority of the experiments, comparisons of diseased plants with healthy plants using Fisher‘s LSD revealed more heavily diseased plants were significantly different from healthy plants. PLS analysis demonstrated the feasibility of detecting early blight infected plants, finding four optimal factors for raw spectra with the predictor variation explained ranging from 93.4% to 94.6% and the response variation explained ranging from 42.7% to 64.7%. The results show the potential of hyperspectral spectroscopy for the detection of early blight in potato plants.(Download PDF) (Export to EndNotes)