Greenhouse Detection Using Aerial Orthophoto and Digital Surface Model

Abstract

Detection of greenhouse areas from remote sensing imagery is important for rural planning, yield estimation and sustainable development. This study is focused on plastic and glass greenhouse detection and discrimination using true color orthophoto and Digital Surface Model (DSM). Initially, the greenhouse areas were detected from true color aerial photograph and the additional normalised Digital Surface Model (nDSM) band using Support Vector Machines (SVM) classification technique. Then, an approach was developed for discriminating plastic and glass greenhouses by utilising the nDSM. The developed approach was implemented in a selected study area from Kumluca district of Antalya, Turkey that includes intensive plastic and glass greenhouses. The obtained results show the effectiveness of the SVM classifier for greenhouse detection with overall accuracy value of 96.15%. In addition, the plastic and glass greenhouses were discriminated efficiently using the developed approach that use nDSM.