Mapping coastal ecosystems and features using a low-cost standard drone: case study, Nayband Bay, Persian gulf, Iran

Abstract

A low-cost standard drone (DJI™ Phantom 4 Pro) has been employed to map the ecosystems and features in coastal areas in the northern Persian Gulf. The Pix4Dcapture® mobile application was selected as a user friendly and simple app to perform an automated and customized drone flight over the selected study area in Nayband Bay, Bushehr province. The flight altitude was selected to be 100 m with 80% overlap for images, which led to the imaging of an area ~ 22 ha (413 × 525 m) in ~13 min (from takeoff to landing the drone). Agisoft™ Metashape PC software was then used to create the orthophoto mosaic from 213 taken photos. Consequently, the ground sampling distance (GSD) of the created orthophoto mosaic was ~3 cm which means it was possible to visually identify the features on the images with a size of 30 cm and more. The orthophoto mosaic was then converted to a topological GIS-based map (in the format of ESRI™ shapefile) by a 2-step procedure including a supervised image classification method coupled with manual editing with an on-screen visual editing method. The final results demonstrated that the overall accuracy of the classified mosaic raster map was 87.6% where the κ (Kappa coefficient) was 0.84. The results also showed that the applied methodology in this study can be used to differentiate the coastal ecosystems and features such as mangrove forests, vegetations, sandy beaches, and deep and shallow water bodies. As a comparison with alternative methods, the cost of implementing drone-based methodology was lower than field surveying and covers a larger area in less time.