An investigation of Triangular Greenness Index performance in vegetation detection

Abstract

Vegetation is one of the most important elements of environment. With the use of satellite systems and especially multispectral images including NIR wavelength, vegetation-oriented studies have increased intensely. With today's advancing technology, besides satellite systems, the unmanned aerial vehicles (UAV) systems are also used to generate similar spectral information being as a platform. However, there are mostly RGB sensors rather than NIR band sensors. More intensive use of RGB sensors has led to the need to use these data in vegetation research. The visible region vegetation index called Triangular Greenness Index (TGI) in the literature is one of the indexes arising from this need. In this study, aerial photographs of an orchard were obtained by the UAV and an orthophoto was produced. Then, the trees detection performance of the TGI in the orchard was compared with the reference data. As a result, the TGI approach can be able to obtain trees with a 95% rate of producer accuracy. This index provides significant support for plant detection studies in the absence of NIR data.