Determination of the pellet parameters by image processing methods

Injection of solid, cryogenic hydrogen isotope pellets in tokamaks is used for particle fuelling as well as for ELM control. The efficiency depends on technical control variables such as pellet size, velocity, frequency and poloidal launch position. Recently developed image processing methods have been improved and adapted in order to evaluate some of these key parameters. An optical flow method has been used for the determination of the ice extrusion velocity based on the image sequences provided by a CCD camera viewing the ice at the exit of the nozzles of the extrusion cryostat. The reconstruction of pellet volume has been performed using images provided by another CCD camera, coupled to a set of optical barriers used for pellet velocity measurement. A Bayesian statistical analysis has been applied, calculating the probability distribution function of the pellet volume based on three measured parameters of the pellet shadow: area, smallest dimension, largest dimension.