ABSTRACT:The purpose of this paper is to analyze how optical pre-processing with polarizing filters and digital pre-processing with HDR imaging, may improve the automated 3D modeling pipeline based on SFM and Image Matching, with special emphasis on optically non-cooperative surfaces of shiny or dark materials. Because of the automatic detection of homologous points, the presence of highlights due to shiny materials, or nearly uniform dark patches produced by low reflectance materials, may produce erroneous matching involving wrong 3D point estimations, and consequently holes and topological errors on the mesh originated by the associated dense 3D cloud. This is due to the limited dynamic range of the 8 bit digital images that are matched each other for generating 3D data. The same 256 levels can be more usefully employed if the actual dynamic range is compressed, avoiding luminance clipping on the darker and lighter image areas. Such approach is here considered both using optical filtering and HDR processing with tone mapping, with experimental evaluation on different Cultural Heritage objects characterized by non-cooperative optical behavior. Three test images of each object have been captured from different positions, changing the shooting conditions (filter/no-filter) and the image processing (no processing/HDR processing), in order to have the same 3 camera orientations with different optical and digital pre-processing, and applying the same automated process to each photo set.

ABSTRACT:Since the advent of the first Kinect as motion controller device for the Microsoft XBOX platform (November 2010), several similar active and low-cost range sensing devices have been introduced on the mass-market for several purposes, including gesture based interfaces, 3D multimedia interaction, robot navigation, finger tracking, 3D body scanning for garment design and proximity sensors for automotive. However, given their capability to generate a real time stream of range images, these has been used in some projects also as general purpose range devices, with performances that for some applications might be satisfying. This paper shows the working principle of the various devices, analyzing them in terms of systematic errors and random errors for exploring the applicability of them in standard 3D capturing problems. Five actual devices have been tested featuring three different technologies: i) Kinect V1 by Microsoft, Structure Sensor by Occipital, and Xtion PRO by ASUS, all based on different implementations of the Primesense sensor; ii) F200 by Intel/Creative, implementing the Realsense pattern projection technology; Kinect V2 by Microsoft, equipped with the Canesta TOF Camera. A critical analysis of the results tries first of all to compare them, and secondarily to focus the range of applications for which such devices could actually work as a viable solution.

-The paper presents an analysis of the 3D data quality generated from small-medium objects by well-known automatic photogrammetry packages based on Structure from Motion (SfM) and Image Matching (IM). The work aims at comparing different shooting configurations and image redundancy, using as high-quality reference the 3D data acquired by triangulation-based laser scanners characterized by a low measurement uncertainty. Two set of tests are presented: i) a laboratory 3D measurement made with the two active and passive approaches, where the image-based 3D acquisition makes use of different camera orientations leading to different image redundancy; ii) a 3D digitization in the field with an industrial laser scanner and two sets of images taken with different overlap levels. The results in the field confirm the relationship between measurement uncertainty and image overlap that emerged in the Lab tests.

No abstract