Introduction

This study aims to explore the use of unmanned aerial vehicles, such as drones, for surveying and mapping purposes. A custom-built drone equipped with an integrated GoPro camera is designed to capture images at predefined geo-locations. As a case study, multiple images of ravine areas and vegetable farms were collected. In order to extract useful information from these images, various softwares have been explored. Amongst them, Pix4D emerged as the standard tool to provide accurate results, which comes with the disadvantage of being very expensive. To overcome this problem, open-source software, such as Meshroom, was also explored, whose results were good, albeit with slightly reduced accuracy. The generated 3D models of the ravine are used to compute area and volume for the prediction of water flow and soil erosion. Moreover, to address the challenges of vegetation analysis, we investigated cost-effective alternatives by leveraging RGB images. Through a comparative analysis of the red, green, and blue channels, we successfully identified areas covered with vegetation, providing valuable insights without the need for expensive multi-spectrum imaging.