Skip to main content
ARS Home » Research » Publications at this Location » Publication #286047

Title: Spectral angle mapper (SAM) based citrus greening disease detection using airborne hyperspectral imaging

item LI, H - University Of Florida
item LEE, W - University Of Florida
item WANG, K - China Agricultural University
item EHSANI, R - University Of Florida
item Yang, Chenghai

Submitted to: International Symposium on Precision Agriculture
Publication Type: Proceedings
Publication Acceptance Date: 9/15/2012
Publication Date: 10/20/2012
Citation: Li, H., Lee, W.S., Wang, K., Ehsani, R., Yang, C. 2012. Spectral angle mapper (SAM) based citrus greening disease detection using airborne hyperspectral imaging. International Symposium on Precision Agriculture. CDROM.

Interpretive Summary: Citrus greening, also known as Huanglongbing (HLB), is a serious citrus disease that causes substantial economic losses to the citrus industry. Early and timely detection is important for the management and control of the disease. This study proposed a novel image classification method for identifying HLB-infected trees using a type of airborne photography known as hyperspectral imagery. Experimental results showed that the new method performed better than two commonly-used image classification techniques. This new technique has the potential for accurate detection of HLB-infected trees, but more research is needed to validate its performance.

Technical Abstract: Over the past two decades, hyperspectral (HS) imaging has provided remarkable performance in ground object classification and disease identification, due to its high spectral resolution. In this paper, a novel method named “extended spectral angle mapping (ESAM)” is proposed to detect citrus greening disease (Huanglongbing or HLB), which is a destructive disease of citrus. Firstly, Savitzky-Golay smoothing filter was applied to the raw image to remove spectral noise within the data, yet to keep the shape, reflectance and absorption features of the spectrum. Then support vector machine (SVM) was used to build a mask to segment tree canopy from the other background. Vertex component analysis (VCA) was chosen to extract the pure endmembers of the masked dataset, due to its better performance compared to other spectral linear unmixing methods. Spectral angle mapping (SAM) was applied to classify healthy and citrus greening disease infected areas in the image using the pure endmembers as an input. Finally, red edge position (REP) was used to filter out most of false positive detections. The experiment was carried out with the image acquired by an airborne hyperspectral imaging system from the Citrus Research and Education (CREC) in Florida, USA. Ground truth including ground reflectance measurement and diseased tree confirmation was conducted. The experimental results were compared with another supervised method, Mahalanobis distance, and an unsupervised method, K-means. The ESAM performed better than those two methods.