Adaptive post-processing of high frame rate image data based on threshold filtering in spectral domian

Abstract

Adaptive post-processing method for high frame rate image data has been developed based on a threshold filtering method in spectral domain to automatically post-process electron cyclotron emission (ECE) imaging data for visualization of magnetohydrodynamics (MHD) instabilities in magnetized high-temperature plasma. This method uses the white noise contained in data having nominally normal distribution to set a threshold criterion for significant power spectral components. The white noise and system noises are deleted by comparing with the threshold value and average spectral power density in the region of interest, respectively. The remaining spectral components are weighted by a sigmoid-like function to the effect of emphasizing MHD components. Examples of ECE images processed by conventional (manual) filtering are compared with automatic threshold filtered images, demonstrating equivalent or improved results in post-processing by the threshold filtering.