Vegetation classification of Coffea on Hawaii Island using Worldview - 2 satellite imagery

Authors: GAERTNER, JULIE - University Of Hawaii; GENOVESSE, VANESSA - California State University; POTTER, CHRISTOPHER - National Aeronautics And Space Administration (NASA); SEWAKE, KEVIN - University Of Hawaii; Manoukis, Nicholas

Submitted to: Journal of Applied Remote Sensing (JARS)
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/26/2017
Publication Date: 10/13/2017
Citation: Gaertner, J., Genovesse, V., Potter, C., Sewake, K., Manoukis, N. 2017. Vegetation classification of Coffea on Hawaii Island using Worldview - 2 satellite imagery. Journal of Applied Remote Sensing (JARS). 11(4):046005. https://doi.org/10.1117/1.JRS.11.046005.
DOI: https://doi.org/10.1117/1.JRS.11.046005

Interpretive Summary: This paper describes two methods for detecting coffee fields using remote sensed image data. Using Worldview-2 satellite imagery the first method, based only on light wavelength information, was able to attain about 69% overall accuracy. The second method, based on both light wavelengths and shapes of objects, was superior, attaining 76% overall accuracy. The methods were used to detect open-air coffee fields in the Kona region of Hawai’i island, USA. The methods described here may be useful for quantifying coffee industries, as well as for preparing for and mitigating coffee pests.

Technical Abstract: Coffee is an important crop in tropical regions of the world; about 125 million people depend on coffee agriculture for their livelihoods. Understanding the spatial extent of coffee fields is useful for management and control of coffee pests such as Hypothenemus hampei (Coffee Berry Borer), other pests that use coffee fruit as a host for immature stages such as Ceratitis capitata (Mediterranean fruit fly), economic planning, and for following changes in coffee agroecosystems over time. We present two methods for detecting Coffea arabica fields using remote sensing and geospatial technologies on WorldView-2 high resolution (2 m) spectral data of the Kona region of Hawaii Island, USA. The first, a pixel based method using a maximum likelihood approach, attained 72% producer accuracy and 69% user accuracy based on analysis of 104 ground truth testing polygons. The second method, an Object-Based Image Analysis (OBIA) approach employing both spectral and textural information, significantly improved accuracy, resulting in 76% producer accuracy and 94% user accuracy for the same testing areas. We conclude that the OBIA method is useful for detecting coffee fields grown in the open, and use it to estimate the distribution of about 1050 hectares under coffee agriculture in the Kona region in 2012.