Northeastern Section - 48th Annual Meeting (18–20 March 2013)

Paper No. 6
Presentation Time: 3:40 PM

PREPARING IMAGE MAPS, PHYSICAL MODELS, AND ANALYZING TOPOGRAPHIC FORM USING TERRESTRIAL LASER SCANNING DATA COLLECTED AT DINOSAUR STATE PARK, ROCKY HILL, CT


HYATT, James A., Environmental Earth Science Department, Eastern Connecticut State University, 83 Windham Street, Willimantic, CT 06226 and ROSIENE, Joel, Mathematics and Computer Science Department, Eastern Connecticut State Univ, 83 Windham Street, Willimantic, CT 06226, hyattj@easternct.edu

Terrestrial Laser Scanning (TLS) is a rapidly evolving technology ideally suited to detailed surface mapping of relatively small sites where there is interest in characterizing topography, building 3D visualizations and physical models, and/or quantifying surface forms. We illustrate these applications by describing ongoing efforts to map, model, and analyze a Eubrontes trackway at Dinosaur State Park (DSP) in Rocky Hill Connecticut using TLS. DSP attracts ≈50,000 visitors annually to view the enclosed trackway surrounding dioramas and displays, and an outdoor arboretum. We used a Trimble VX spatial station to collect a ≈380,000 survey point cloud for 125 m2 of the trackway. Data were collected from 4 scanning stations georeferenced using resection points on the surrounding dioramas. The point cloud was segmented, decimated, and slightly smoothed to generate a 3D mesh of the trackway. Georeferenced VX imagery was used to texture the mesh as a first step in preparing educational maps and to isolate forms. Bright side-lighting at the trackway, however, degraded VX imagery. Accordingly, low-level (8 m), near-vertical pole-based DLSR images (50 mm and 35 mm focal lengths) were captured and used to retextured the mesh prior to orthorectification and preparation of photo-maps. We also experimented with extracting and processing data from a single track to prepare physical models using on-line 3D printing. Our ongoing form analyses extract feature data from selected tracks with the intent of testing whether one set of tracks can be distinguished from another. Data preparation included isolating selected forms, normalizing coordinate space for each based on best-fit planes and centroids, and exporting individual form point clouds for analysis. 3D feature data are derived from truncated track fill volumes, whereas planimetric slices were used to define 2D feature properties. Form boundaries expressed in polar coordinates relative to centroids were parameterized by boundary arc length in order to unwrap forms and calculate Fourier shape descriptors (FSD). FSD’s together with Gaussian curvature texture maps are used to quantify and compare forms. Although analyses are ongoing they illustrate the degree to which high resolution digital terrain data promotes cross-disciplinary analysis of surface landscapes.