Citrus Greening Disease Detection Using Airborne Multispectral And Hyperspectral Imaging
1C. Yang, 2L. G. Albrigo, 3, 2W. Lee, 2A. Kumar, 2R. Ehsani
1. USDA-ARS
2. University of Florida
3.
Citrus greening disease (Huanglongbing or HLB) has become a major catastrophic disease in Florida’s $9 billion citrus industry since 2005, and continued to be spread to other parts of the U.S. There is no known cure for this disease. As of October 2009, citrus trees in 2,702 different sections (square mile) in 34 counties were infected in Florida. A set of hyperspectral imageries were used to develop disease detection algorithms using image-derived spectral library, the mixture tuned match filtering (MTMF), spectral angle mapping (SAM), spectral feature fitting (SFF), and spectral analyst tool in the hyperspectral imaging software. However, due to positioning errors in ground truthing data, there were many false positives.
A new set of multispectral and hyperspectral imagery will be taken in early December 2009 from different imaging sites having citrus trees of various degrees of infection in southern Florida. The multispectral sensor provides images with 2048 by 2048 pixels in blue, green, red, and near-infrared (NIR) spectral bands. Actual size of a pixel on the ground will be 0.14-1.12 m. The hyperspectral sensor is configured to record imagery with 128 bands from blue to NIR spectral ranges (457.2 to 921.7 nm at 3.63 nm intervals). The imagery has a swath width of 640 pixels and a radiometric resolution of 12 bits. The pixel sizes will be 1-3 m. The previously developed algorithms will be applied to this new set of images and results will be reported.
