Monitoring of young forest restoration plantations from drone-based aerial images and deep learning

Abstract

Remote sensing tools, such as Piloted Aerial Systems Remotely (RPAS) coupled to RGB (Red-Green-Blue) sensors, they feature High efficiency in capturing data and high-resolution images. So neural networks are configured as practical instruments for the automation and optimization of processes. The central purpose of the work is to test the application of RPAS images in conjunction with machine learning algorithms to improve the monitoring of ecological indicators in the restoration Forest. The project will be developed in distinct phases, covering data collection, processing of flight information and implementation of the classification algorithm. This methodological framework aims to achieve a comprehensive and effective analysis of the ecological indicators related to forest restoration. The execution of these procedures will take place for the analysis of images from the experimental areas of the project "Comparative Analysis of Active Restoration Techniques involving Direct Sowing and/or Seedling Planting" (FAPESP 2023/01325-2) and the experiment MataDIV. This strategic approach aims to optimize time and resources, providing the Efficient application of remote sensing and artificial intelligence tools mentioned, which can contribute to future monitoring initiatives in the future. scale in ecological restoration.