Quantifying the density-quality of photogrammetrically created point-clouds of linear architectural/urban elements as a function of shooting distances and number of camera positions I.E. shooting-directions

Abstract

This paper examines the impact of various object-to-camera distances and the number of station-points i.e. various shooting directions with regard to the obtained Density-quality of photogrammetrically created Point-clouds-as digital representations of the existent linear architectural/urban objects/elements. According to an artificial (purified) experimental scene used, the conclusion is that with the chosen focal lengths/object-to-camera distances, with shooting directions perpendicular to the axis of that object, with station-points uniformly radially distributed around it (at a circle of 360deg), and with the obtained values of photogrammetric-software process-quality outputs which belong to the recommended ranges, the achieved density-level of the created Point-clouds may be treated as independent on the camera's radial-movement angle but dependent on the percentage of "Object's Photo-Coverage": the lower the Coverage, the lower the density. Also, regardless of the Coverage level, the majority of the generated points are generally more "densimetrically" precise than they are "densimetrically" accurate.

Point-clouds as photogrammetric representations of linear objects surfaces: The impact of extrinsic shooting parameters on the change of roughness-quality of their textures

“…Since the nature of this experiment is identical to that of the initial experiment [2] (herewith: "IE"), the test-field setup is directly inherited from the IE (Figures 1(a) and 1(b)). This fact relates to the applied artificial light source, used experimental vertically positioned object /stick (L=25cm, Diameter=10mm / Figure 2 Hence, there are station-points placed at three different horizontal circle-paths, with radii of 70 cm, 110 cm and 220 cm -varying in number: from 24 (determined by 15deg camera radial-movement angle), via 12 (determined by 30deg radial movement) to 8 (determined by 45deg radial movement).…”

“…The previous research in this field was focused on investigating several tasks that refer to: -comparative analysis of achieved accuracy of two terrestrial/close-range photogrammetry processes: semi-automatic and automatic, based on analysis of deviation of distances between control points defined on a chosen 3D-object from distances between their digitalized representations [12], -finding an efficient reverse engineering method /algorithm which can be used for precise and accurate recognition of surfaces from corresponding structured point-clouds, when those surfaces are primarily represented by elliptical segments (consisted of elliptical cylinders and ellipsoids, including spheres, rotating cylinders and cones) but, also, represented by higher-degree surface-segments, such as elliptical tori [10], -quantifying the accuracy of dense surface (and terrain) modelling (herewith: "DSM") obtained using a broad spectrum of photogrammetric packages [4,6,8,9], -quantifying the metric quality of unstructured linear objects point-clouds -obtained for horizontal shooting directions as a function of object inclination angles [1], -quantifying the density quality of unstructured linear objects point-clouds -obtained for horizontal shooting directions as a function of object inclination angles [1], -quantifying the metric quality of unstructured linear objects point-clouds -obtained for shooting directions perpendicular to those objects as a function of shooting-distances and number of camera positions shooting-directions [2], -quantifying the density quality of unstructured linear objects point-clouds -obtained for shooting directions perpendicular to those objects as a function of shooting-distances and number of camera positions shooting-directions [3], and -comparing the accuracy of professional and consumer grade 3D-printers in complex models production [13], that enables all results of the previously mentioned research to be synergically used for precise and accurate 3D processing of linear objects surfaces in the whole (complying thus with complex needs of contemporary engineering praxis).…”