USING REMOTE SENSING AND GIS FOR DETECTING AND MAPPING INVASIVE WEEDS IN RIPARIAN AND WETLAND ECOSYSTEMS
Title: Citrus greening disease detection using airborne multispectral and hyperspectral imaging
Submitted to: International Symposium on Precision Agriculture
Publication Type: Proceedings
Publication Acceptance Date: August 27, 2010
Publication Date: September 25, 2010
Citation: Kumar, A., Lee, W.S., Ehsani, R., Albrigo, L.G., Yang, C., Mangan, R.L. 2010. Citrus greening disease detection using airborne multispectral and hyperspectral imaging. International Symposium on Precision Agriculture. 2010 CDROM.
Interpretive Summary: Huanglongbing (HLB), also known as citrus greening, is a serious citrus disease that affects all citrus cultivars and causes rapid deterioration of citrus trees. Early detection is very critical for the management and control of the disease. This study evaluated high resolution airborne multispectral and hyperspectral imagery for identifying HLB-infected citrus trees in Florida. Several image analysis techniques were used to separate infected trees from healthy trees based on the imagery. Preliminary results showed that high resolution imagery in conjunction with image analysis techniques can identify infected trees with the yellowing symptom, but early detection remains to be a challenge.
Hyperspectral imaging can provide unique spectral signatures for diseased vegetation. Airborne hyperspectral imaging can be used to detect potentially infected trees over a large area for rapid detection of infected zones. Ground inspection and management can be focused on these infected zones rather than entire grove, making this less labor intensive and time consuming. This paper proposes a method to detect the citrus greening disease (Huanglongbing or HLB) infected areas in citrus groves using airborne hyperspectral and multispectral imaging. This would prevent further spread of infection, if complimented by development of efficient management plans of infected areas. Hence, airborne hyperspectral imaging would provide much faster results over a wide range of area. An aerial hyperspectral image with a spectral range of 457-921 nm spanning across 128 spectral bands was acquired from an HLB infected citrus grove in Florida in December 2009. The imagery had a 3.6 nm spectral resolution and 1 m spatial resolution. A multispectral image with 4 bands (red, green, blue, and NIR) spanning across a spectral range of 480-830 nm with 40 nm bandwidth was also acquired from the same grove. Polymerase chain reaction (PCR) test based ground truthing of this area had been carried out where the infected tree coordinates were recorded. An image derived spectral library was built using the above images and categories of Healthy and HLB infected pixels were created based on the PCR results and locations of the infected trees. Ground measurements were obtained for Healthy and HLB infected citrus trees from the same grove along with their degrees of infection. A hyperspectral imaging software (ENVI, ITT VIS) was used for the analysis of these images. HLB infected areas were identified using image-derived spectral library, the mixture tuned match filtering (MTMF), the spectral angle mapping (SAM), and linear spectral unmixing (LSU). It was seen that accuracy of MTMF method was greater than the other methods. The accuracy of SAM using multispectral images was comparable to the results of the MTMF and also gave higher accuracy when compared to SAM analysis on hyperspectral images.