A proposal of a new automated method for SfM/MVS 3D reconstruction through comparisons of 3D data by SfM/MVS and handheld laser scanners

Abstract

SfM/MVS photogrammetry has received increasing attention due to its convenience, broadening the range of its applications into archaeology and anthropology. Because the accuracy of SfM/MVS depends on photography, one important issue is that incorrect or low-density point clouds are found in 3D models due to poor overlapping between images. A systematic way of taking photographs solve these problems, though it has not been well established and the accuracy has not been examined either, with some exceptions. The present study aims to (i) develop an efficient method for recording pottery using an automated turntable and (ii) assess its accuracy through a comparison with 3D models made by laser scanning. We recorded relatively simple pottery manufactured by prehistoric farmers in the Japanese archipelago using SfM/MVS photogrammetry and laser scanning. Further, by measuring the Hausdorff distance between 3D models made using these two methods, we show that their difference is negligibly small, suggesting that our method is sufficiently accurate to record pottery.

Macro-Scale Population Patterns in the Kofun Period of the Japanese Archipelago: Quantitative Analysis of a Larger Sample of Three-Dimensional Data from Ancient Human Crania

“…Three types of laser scanners (i.e., AMETEK Creaform HandySCAN BLACK, HandySCAN BLACK™ | Elite (Berwyn, PA, USA), and Shining 3D Einscan Pro HD (Hangzhou, China)) and SfM/MVS, based on two-dimensional photos, were used to collect 3D data. Previous research has confirmed that 3D models reconstructed using these methods do not exhibit significant differences and maintain fidelity to the original crania [28][29][30]. Given that the original 3D models consist of numerous meshes, making subsequent analyses challenging, we reduced the number of meshes to 1% to 10% of the original using Meshlab [31].…”

Brain MRA 3D Skeleton Extraction Based on Normal Plane Centroid Algorithm

“…Then, the FLANN matching algorithm [25] matches the feature points to different images. Finally, the traditional 3D reconstruction algorithm -SFM algorithm [26] is used to reconstruct the above-processed photos, which can be divided into incremental, global, hybrid, hierarchical, and semantic-based methods of the SFM algorithm [30][31][32]. A hierarchical algorithm is used for image 3D reconstruction, as shown in Figure 6.…”

Experiences of TLS, terrestrial and UAV photogrammetry in Cultural Heritage environment for restoration and maintenance purposes of Royal Racconigi castle, Italy

“…In particular, Structure from Motion (SfM) approach has become quite popular in Close Range Photogrammetry (CRP) thanks to ability to determine the parameters of external orientation without any a priori knowledge of the approximate positions for cameras and 3D points. SfM technique requires, in order realizing 3D models, a block of images with a high degree of overlap that capture the complete 3D structure of the scene viewed from several positions [2]. In addition, using Multi-View-Stereo (MVS) algorithm, it is possible increasing the density of the point cloud generated in SfM process.…”