HawkEye: Automatic Stitching of Hand-Held LIDAR Scans using Photogrammetry and Structure-from-Motion

Abstract

Numerous applications—from aligning mirrors on the latest telescopes to accurate body analysis of burn patients—depend on easy and accurate measurements. The field of metrology specializes in approaches to solve the precise measurement problem. Photogrammetry, in particular, uses targets (either non-coded or coded) placed on or around an object to compute the target positions in three-dimensional (3D) space from two-dimensional (2D) photographs. This time intensive process, although accurate, traditionally results only in the 3D geometry of the target positionsi. Other metrology techniques, such as LIDAR, provide dense 3D geometry of the imaged scene. Unfortunately, such systems are expensive, cumbersome, and time-consuming. Moreover, globally stitching LIDAR scans together is a semi-manual process that relies on accurate user-selected matches or careful calibration of the equipment, or that employs high-precision motorized controls to set device poses. To combat the issues and costs associated with current approaches and processes, we developed a handheld consumer-grade optical scanner and supporting application suite, called HawkEye™. Our system utilizes recent advancements in hybrid photogrammetry/structure-from-motion technologies ii to provide a fully automated stitching process that is inexpensive and provides an accurate and dense reconstruction. 2. Methodology and Results The portable hand-held 3D scanner contains an HD RGB camera and IR time-of-flight sensor.