Abstract

The rapid proliferation of consumer small unmanned aerial systems (sUASs) has expanded ownership to include amateurs and professionals alike. These platforms in combination with numerous open source and proprietary applications tailored to gather aerial imagery and generate 3D point clouds and meshes from aerial imagery, have made 3D modeling available to anyone who can afford an entry-level sUAS. These flight plans force the sensor to remain at greater distances from their targets, resulting in varying spatial resolution of sloped surfaces. The work described here explains the development of a variety of 3D automated flight plans to provide vantage points not achievable by constant-altitude, nadir-looking imagery. Specifically, the issue of roof inspection is addressed in detail. This work generates an automated flight plan that positions the sUAS and orients its sensor such that the focal plane array is parallel to the roof plane based on a priori knowledge of the roof's geometry, greatly reducing single- or two-point perspective. This a priori knowledge can come from a variety sources including databases, a site survey, or data extracted from an existing point cloud. Still images or video from orthogonal flight plans can be used for visual inspection, or the generation of dense point clouds and meshes. These products are compared to those generated from nadir imagery. This novel flight planning approach permits the aircraft to fly the orthogonal flight plans from start to finish without intervention from the remote pilot. This work is scalable to similar sUAS-based tasks including aerial-based thermography of buildings and infrastructure.

Publication Date

7-17-2017

Document Type

Thesis

Student Type

Graduate

Degree Name

Imaging Science (MS)

Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)

Advisor

Carl Salvaggio

Advisor/Committee Member

Jan Van Aardt

Advisor/Committee Member

Jeff B. Pelz

Campus

RIT – Main Campus

Share

COinS