Simulation of Photogrammetry-Based 3D Data Acquisition

Abstract

The paper presents a simulation approach to photogrammetry-based three-dimensional (3D) data acquisition. Photogrammetry requires capturing of series of overlapping photos with certain properties from which 3D reconstruction is later obtained. Scanning a building or a human or jewellery requires different numbers of cameras, setup parameters, spatial orientations, etc. Without precise information on how to effectively take photos, obtaining them can be tedious work without any guarantees that it will provide sufficient 3D reconstruction quality. The proposed simulation approach aims to ease the aforementioned burdens and contributes by improving the process of photogrammetry-based 3D data acquisition. The presented simulator is tested in the context of the development of a 3D scanning system for human body scanning and avatar creation. The experiments confirm that the proposed method leads to an improved quality of 3D object reconstruction in comparison to previous practice in the field of 3D human scanning. Further, it lowers the cost and shortens the time required for the industrial process of construction of 3D scanning systems, thus confirming the value and validity of the presented approach.

Research on the Design and Automatic Recognition Algorithm of Subsidence Marks for Close-Range Photogrammetry

“…s is an arbitrary scale factor. M is the matrix of the camera intrinsic parameters as shown in Formula (11):…”

“…is the matrix of the camera intrinsic parameters as shown in Formula (11): = 0 0 0 0 1 (11) where and are the focal distances in pixels. The physical focal distance of the camera in millimeters is fixed, while the shape of each picture element is not a square but a rectangle, so there are two focal distances; one is in x direction and the other is in y direction.…”

“…There are many advantages to close-range photogrammetry [10][11][12]. The first is that a large number of geometry structure information of the objects to be measured can be acquired instantly.…”

Comparative evaluation of keypoint detectors for 3d digital avatar reconstruction

“…To capture image data, we use multi-view stereo setup with 54 high-resolution RGB calibrated cameras, conceptually similar to the one described in [39] and shown in Fig. 2.…”

“…2. During the image acquisition process, the subject is standing in a center of the setup with legs slightly apart and arms positioned at an approximately 30-degree angle away from the body (so-called A-pose) [39]. Fig.…”