Estimativa do volume de pilhas de madeira utilizando dados de aeronave remotamente pilotada de baixo custo

Abstract

The concepts of Innovation and Control are frequently applied in current production processes, including the forestry sector. Efficient wood stock management is an important task for forestry companies, ensuring a continuous flow and appropriate logistics for operations that connect forestry to industry. The quantification of stacked wood volume is essential for this management, requiring the collection of field data, such as log length, pile length, and pile height. In several regions of the world, manual measurement of these dimensions remains common, making the process labor-intensive and time-consuming, which complicates consistent stock control. However, the use of aerial surveys in forestry management has grown, driven by technical advances in remotely piloted aircraft (RPA) and the development of processing algorithms based on structure from motion (SfM) and multiview stereopsis (MVS). Nevertheless, there is still a need to evaluate the variables that influence stacked wood volume estimates with aerial surveys, including the flight plan itself, which represents an operational gap that warrants attention in forestry applications research. Thus, the present study aims to estimate the volume of stacked wood in the field using remote sensing data collected by a low-cost drone. For this purpose, a preliminary analysis was conducted to identify the most appropriate flight configuration for estimating the volume of a wood stack, varying the overlap rate and camera angle relative to the aircraft’s axis. Based on the results of this analysis and existing literature, a new field evaluation was conducted to verify whether different flight plans would yield accurate estimates of wood pile volumes. Both tests were performed using low-cost equipment and without the use of control points. The results obtained in this study showed that aerial surveys with low-cost drones and without the use of control points are capable of providing accurate volumetric estimates of stacked wood, contributing to advancements in drone applications within the forestry sector, particularly in forest management.