Thermal / Visible Autonomous Stereo Visio System Calibration Methodology for Non-controlled Environments

Abstract

This paper presents a fast and low cost methodology for thermal / visible stereo vision system calibration with high reliability for non-controlled external environments. The proposed methodology uses a non-powered calibration rig to map the coordinates, field of view and magnification distortions from visible and thermal vision and a self-calibration analysis procedure to estimate metrological uncertainty of imaging correspondence. Tests were conducted to determine the effectiveness of the developed calibration rig. 

Herramienta software para inspección termográfica usando fusión multimodal de imágenes térmicas y de espectro visible

“…(Fluke, 2015b) FLIR Tools (Fluke, 2015a) DuoVision (Trotec, 2015) IR Analizer (Infrared, 2015) IrSoft (Testo, 2015) TermUV (Benitez et al, 2005) The calibration of a hybrid camera system might be thought similar to stereo camera rig calibration problem. However, according to Ursine et al (2012) it is not enough to know the relative translation and rotation between the two cameras, since other important problems must be considered such as field of view, sensor size, and resolution. As observed in Table 1, the intrinsic calibration process is performed in most cases using the method proposed in (Bouguet, 2013); nevertheless, all works show different approaches for the extrinsic calibration process.…”

“…Another important aspect to classify these works is the calibration pattern used in Prakash et al (2006), StLaurent et al (2010), Senya et al (2011) andYang &Chen (2011), in which active thermography was used to build the calibration pattern. In contrast, in Ursine et al (2012) and , a calibration pattern which exploits the different emissivity values of different materials was used.…”

“…The authors use this setup in order to recover 3D thermal and visual information, however, the working distance is short and not suitable for outdoors thermography. The work of Ursine et al (2012), has two contributions, a brute force pixel to pixel alignment method in order to obtain the extrinsic parameters and the calibration pattern; and the calibration pattern design. This calibration pattern was implemented in this paper.…”

“…It is worth noting that the base line between cameras was as short as possible in order to achieve the highest FOV overlaps. Figure 1b shows the calibration pattern built considering the work of Ursine et al (2012). The calibration pattern was constructed using two different materials with different emissivity such as wood and copper.…”

Passive calibration board for alignment of VIS-NIR, SWIR and LWIR images

“…Ursine et al [4] use a checkerboard pattern made of copper squares of very low emissivity above a background painted with a high emissivity black ink to calibrate a pair of color and IR cameras. They obtained interesting results for outdoor scenes, but they need to take precautions to minimize specular reflections on their extremely low emissivity (0.09) copper material.…”