Photogrammetry and Machine Vision

Abstract

We focus on sensors, systems, processing methods and applications typically used for high accuracy image-based measurements, as e.g. in industrial quality control. We cover also motion tracking, as applied in biomechanics and animation. We aim at a deeper understanding of refined sensor and system modeling, estimation procedures, statistical testing. A project is carried through the semester.

Objective

Understanding the methods of analytical and digital close-range photogrammetry and machine vision. Experiences with and critical analysis of image-based and other sensors and measurement techniques used in this domain. Familiarization with systematic error modeling, blunder detection, statistical testing, and profound accuracy analysis of photogrammetric systems and results. Experience with competent modeling procedures for sensors, systems and processing algorithms. Understanding of the power and pitfalls of image-based high-accuracy measurement techniques. Critical evaluation of methods and important products of terrestrial applications: Industrial quality control and process monitoring, motion tracking for biomechanics and animation (e.g. in the movie industry) and others.

Content

Definitions, applications, markets, foundations, history. Photogrammetric network, measurements and adjustment: digital cameras (CCD, CMOS, high-speed, video) and digital images, measurements (coded targets, LSM), bundle adjustment, self-calibration, statistical analysis, precision criteria, reliability theory, image network configuration, surface generation, texture mapping, visualization. Active sensors: measurement principle, laser light, structred light, system calibration, registration, TOF cameras, applications. Panoramic cameras: principles, techniques, sensor modeling, applications. Movement analysis: particle tracking, motion capture, crash test, deformation analysis.

Resources