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.
Detecting Marssonina Blotch Using Hyperspectral Imaging and Hierarchical Clustering
Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.orgCitation: 2015 ASABE Annual International Meeting 152150680.(doi:10.13031/aim.20152150680)
Authors: Brianna B Posadas, Won Suk Lee, Youngki Hong, Sangcheol Kim
Keywords: apples, Asia, hyperspectral imagery, near-infrared reflectance, spectral analysis.
Abstract. Apple Marssonina blotch disease (AMB) is caused by the fungus Diplocarpon mali. Once an apple tree is infected, 40 days are required for the disease to be detected visually. The disease symptoms begin on the leaves as small brown lesions followed by foliar chlorosis. The trees defoliate prematurely, reducing the number of apples that particular tree can produce that year and considerably reducing profits. Countries where apple blotch disease is reported as a serious problem include the Republic of Korea, India, and China. China is the largest producer of apples with a production of over 37,000,000 metric tons. Once an orchard has been infected with D. mali, the only control measure is to cut down and remove the diseased tree. Preventative measures include spraying with fungicide; however, D. mali has been shown to have low sensitivity to copper fungicides and is becoming resistant to thiophanate-methyl. This project utilizes near-infrared (NIR) spectral reflectance to detect AMB. NIR has been used to detect various plant contents that affect health, such as moisture, nitrogen, and other nutrients, and has been demonstrated to be a promising early detection method for various plant diseases. Spectral data, taken with a hyperspectral camera, was used to develop an algorithm that can be calibrated as an early detection of AMB. Using the feature selection method of hierarchical clustering on images of Fuji apples, it was shown that it is a plausible method that can be developed into a robust early detection models.(Download PDF) (Export to EndNotes)