Tsallis Entropic thresholding based segmentation of Gas Puff Images of Plasma using Normalize Grey Level Spatial Correlation Histogram
Bhargav Shah, Raju Daniel · 2020
Edge turbulence in fusion reactor leads to loss of confinement of plasma which is used to generate energy from the fusion reactor if confined within. Gas Puff Imaging (GPI) diagnostic is used to identify the characteristic of edge turbulence using ultra high speed camera. In this paper, Normalized Gray Level Spatial Correlation (NGLSC) Histogram based image segmentation technique using Tsallis Entropy is presented to segment the edge turbulence from the Gas puff Images. It generates the 2D histogram using pixel's gray level and normalize local gray level. It is then considered to locate the optimum threshold using Tsallis entropy to segment the image into background and edge turbulence. The proposed technique is compared with the existing 1D and 2D Entropy based Image segmentation techniques. The result obtained shows that the intended technique gives better result in terms of Uniformity Values compared to the current standard technique 1D and 2D Entropy.