Larsen Hannah J, Bennett Matthew R
Department of Life and Environmental Sciences, Faculty of Science and Technology, Bournemouth University, Poole, BH12 5BB, U.K.
J Forensic Sci. 2020 Sep;65(5):1722-1729. doi: 10.1111/1556-4029.14455. Epub 2020 May 14.
This paper examines the reliability of Structure from Motion (SfM) photogrammetry as a tool in the capture of forensic footwear marks. This is applicable to photogrammetry freeware DigTrace but is equally relevant to other SfM solutions. SfM simply requires a digital camera, a scale bar, and a selection of oblique photographs of the trace in question taken at the scene. The output is a digital three-dimensional point cloud of the surface and any plastic trace thereon. The first section of this paper examines the reliability of photogrammetry to capture the same data when repeatedly used on one impression, while the second part assesses the impact of varying cameras. Using cloud to cloud comparisons that measure the distance between two-point clouds, we assess the variability between models. The results highlight how little variability is evident and therefore speak to the accuracy and consistency of such techniques in the capture of three-dimensional traces. Using this method, 3D footwear impressions can, in many substrates, be collected with a repeatability of 97% with any variation between models less than ~0.5 mm.
本文探讨了运动结构（SfM）摄影测量法作为一种获取法医鞋印工具的可靠性。这适用于摄影测量免费软件DigTrace，但同样适用于其他SfM解决方案。SfM仅需要一台数码相机、一个比例尺以及在现场拍摄的一系列有关该痕迹的倾斜照片。输出结果是该表面及其上任何塑性痕迹的数字三维点云。本文第一部分研究了在同一痕迹上反复使用摄影测量法时获取相同数据的可靠性，而第二部分评估了不同相机的影响。通过测量两点云之间距离的云对云比较，我们评估了模型之间的变异性。结果突出显示出明显的变异性非常小，因此说明了此类技术在获取三维痕迹方面的准确性和一致性。使用这种方法，在许多基质中，3D鞋印的采集重复性可达97%，模型之间的任何差异小于约0.5毫米。
