A New Vegetation Index in Short-Wave Infrared Region of Electromagnetic Spectrum

Abstract

Vegetation index algorithms based on radiance and/or reflectance data in the Visible Near Infrared (VNIR) band are created for multispectral and hyperspectral images to detect vegetation. By this way, vegetation and soil regions can be separated from each other. Furthermore, vegetation maturity and health can be observed and tracked. Besides VNIR band cameras, there are many cameras working in the Short Wavelength Infrared (SWIR) band. Therefore, the plant areas should also be detected in SWIR band images. For the purpose of the vegetation detection, a new vegetation index algorithm is created for SWIR hyperspectral images in this research. The proposed method’s feasibility and precision is evaluated using SWIR band hyperspectral data collected at various locations and times during the year. Test results show that besides vegetation regions, soil and water regions in the SWIR band hyperspectral images can be detected successfully. On Hyperion hyperspectral data, 98.9% and 98.5% mean scores are achieved for recall and precision, respectively.

Keywords

• Vegetation index
• short wave infrared
• hyperspectral image
• precision agriculture