Calibration and applications of visible imaging cameras on the C-2U advanced beam-driven field-reversed configuration device

REVIEW OF SCIENTIFIC INSTRUMENTS 89, 10E103 (2018)
Calibration and applications of visible imaging cameras on the C-2U
advanced beam-driven field-reversed configuration device
E. M. Granstedt,a) D. Fallah, M. C. Thompson, and TAE Teamb) TAE Technologies, Inc., 19631 Pauling, Foothill Ranch, California 92610, USA
(Presented 17 April 2018; received 24 April 2018; accepted 5 June 2018; published online 10 August 2018)
Two filtered, fast-imaging instruments, with radial and axial views, respectively, were used on the C-2U device to visualize line emission from impurities and hydrogenic neutrals. Novel calibration techniques needed to be developed for these instruments because the accelerated pace of C-2U opera- tions precluded access to the interior of the vacuum vessel, targets used in typical calibration methods were not available, and in order to account for effects which have not been sufficiently addressed in the literature. Spatial calibration involved optimizing parameters in a generic camera model: ex situ using a checkerboard target and in situ using the vacuum vessel port geometry. Radiometric calibration was performed ex situ in three stages. First, the camera relative response function was mapped using an algorithm developed for high dynamic-range imaging. Second, the non-uniformity of the optical sys- tem was measured using a large LCD monitor with a characterized angular emission pattern. Finally, the absolute photon efficiency of each interference filter was determined using a calibrated uniform radiance source while also accounting for reduction in the filter transmission for off-normal rays. Periodically during the run campaign, line emission from neutral beams fired into a gas target was used as an in situ reference to check for degradation of viewport transmission. One application using calibrated camera data was tomographic reconstruction of passive impurity emission, which provided
a sanity check to the excluded-flux radius inferred from
by AIP Publishing. https://doi.org/10.1063/1.5037453
I. INTRODUCTION
High-speed cameras equipped with narrow-band inter- ference filters have been used on fusion plasma experiments to visualize numerous phenomena including the global dis- charge evolution, edge instabilities,2 plasma-material interac- tion,3 and particle fueling.4 To best observe these phenomena and enable 3D tomographic reconstruction, two camera sys- tems with approximately orthogonal views were fielded on the C-2U1 device: one wide-angle radial view and one nearly axial view (Fig. 1). Each instrument consisted of a high-speed camera, a 10-position filter wheel with position feedback, a re-entrant sapphire viewport, and an optical periscope con- sisting of discrete lenses. The entire system is mounted on a bellows actuator so that the viewport could be retracted behind a gate valve when titanium evaporators were energized. The radial camera was a monochrome Sanstreak Edgertronic SC1, and the axial camera was a Vision Research Phantom v5.2. A detailed description of the instruments can be found in Ref. 5.
High-speed imaging yields a trove of data containing qual- itative information about the plasma evolution and stability. In order to address quantitative questions, however, such as the density of neutrals or impurities, the instruments must be spa- tially and radiometrically calibrated. Spatial calibration allows
Note: Paper published as part of the Proceedings of the 22nd Topical Confer- ence on High-Temperature Plasma Diagnostics, San Diego, California, April 2018.
a)Electronic mail: egranstedt@tae.com
b)TAE Team members are listed in H. Gota et al., Nucl. Fusion 57, 116021 (2017).
wall-mounted magnetic sensors. Published
each camera pixel to be mapped to a sight-line through the plasma, while radiometric calibration maps detector counts to photon radiance.
The operations’ schedule for the C-2U program prohibited in-vessel access for calibration of these systems, and large-area calibration targets were not available. As a result, novel cali- bration techniques had to be developed. This paper describes the calibration of the instruments and a sample application for the calibrated data.
II. SPATIAL CALIBRATION
Each pixel of the camera detector collects light from an emitting volume determined by the position and orientation of the optical system and its components. These cumulative effects are modeled using a camera model. Once the parame- ters of the model are determined using a calibration procedure, each pixel can be mapped to a “sight-cone” through the vac- uum vessel. The “generic camera model”6 used in this work assumes an infinitesimal aperture (vanishing point) with finite depth-of-field and diffraction effects neglected. The model consists of four parts: an external calibration corresponding to the position and orientation of the vanishing point, a radially symmetric mapping function which maps distances in physi- cal space to pixel space (and accounts for barrel/fisheye and pincushion distortion), a function which accounts for asym- metric (e.g., tangential) distortion of the optical system, and an internal calibration matrix which accounts for detector misalignment relative to the optical axis. In principle,
0034-6748/2018/89(10)/10E103/5/$30.00 89, 10E103-1 Published by AIP Publishing.