Low Cost Target Design and Detection for Camera Calibration in Image Based Close Range Inspection Applications

Abstract

Camera calibration is a widely covered and wellknown topic, which has been discussed and well described in many publications. However, the detection of the corresponding calibration markers can be problematic if the imaging setup considers a very short working distance, a huge magnification and special lighting setups, which is the case in almost every industrial inspection system, especially in surface inspection. Low-cost calibration targets can have poor quality in terms of surface homogeneity, resulting in reflection and illumination artifacts. In this paper, we present a distance transform based approach for detecting calibration markers of those targets. As an example application, the typical setup of a surface inspection system for printed circuit boards is used. This approach ensures a built-in calibration process without the need of changing the imaging setup.