Component Contour-Based Viewpoint Planning for UAV Scene Coverage

In recent years, drones have been widely used for monitoring natural environments and for the three-dimensional (3D) reconstruction of infrastructure. The growing use of drones is the result of technological advances in drone equipment and data processing. However, drones face challenges with efficiency and path planning during 3D reconstruction and mapping, especially when they must navigate complex object surfaces that are divided into various parts. These surfaces typically feature intricate details that are vital for the precision and faithfulness of 3D models. Capturing these details is essential for the quality of reconstruction outcomes. To address this issue, we are introducing a novel method for generating viewpoints that specifically considers the segmented sections of objects. This method uses the contours of the object surfaces to direct the sampling process. Our goal is to establish a compact set of viewpoints that ensures comprehensive coverage for each component, thus capturing the intricate details of each surface effectively. Using contours, our method fine-tunes the selection of viewpoints to minimize redundancy and maximize the capture of information for each segmented piece. Through a series of simulation experiments, we have validated that our method achieves superior coverage with a reduced computational time for the segmented components of the scene.