Evaluation of plant reflectance response with elevation using multispectral images captured by an unmanned aerial vehicle (UAV)

Abstract

The survival, development and productivity of plants can be affected by elevation. Remote sensing has been used to study altitudinal gradient and plant reflectance. Plant reflectance is an important factor for determining plant health and phenology. This study presents a technique to support a better understanding of how plant reflectance is associated with elevation. In particular, this study determined the effect of elevation on reflectance of pineapple. This study was conducted at Polomolok, South Cotabato, Philippines. The Unmanned Aerial Vehicle (UAV) platform, eBee Ag, onboard the Parrot Sequoia multispectral camera was used to capture multispectral images at 121 meters flight altitude with 80% image overlap on eight areas located at 400-500 meter-above-sea-level (masl) (Location 1) and 650-700 masl (Location 2) elevations. Image stitching was done through Pix4DMapper 3.1 using Ag Multispectral template. The root mean square error (RMSE) for the x-, y- and z- direction justified good and comparable accuracy for all images stitched. Multispectral images captured by an UAV could discriminate plant reflectance response in different elevations. Most of the data demonstrate a moderate positive correlation between elevation and green, red, red-edge and near-infrared reflectance. The only exceptions were correlations between elevation and red-edge reflectance (no correlation), and between elevation and near-infrared reflectance (weak correlation) in Location 2.