Fast Recognition of Bright Spot Structures in Divertor Region Based on Improved DeepLabv3+ Image Segmentation on EAST

Processes, Journal Year: 2025, Volume and Issue: 13(3), P. 899 - 899

Published: March 19, 2025

The presence of a bright spot structure in the divertor region during discharge process, indicative localized overheating, has been observed through multi-band and high-speed endoscope diagnostic on Experimental Advanced Superconducting Tokamak (EAST). This deposition hyperthermal heat flux can lead to erosion melting target plate material, thereby posing significant risk safe operation device. Moreover, it may introduce impurities into main plasma, negatively impacting plasma performance. Therefore, real-time monitoring rapid identification overheating regions experiments are crucial. In this context, paper proposes an improved DeepLabv3+-based highlight image-segmentation algorithm, which uses minimum value, image difference method, Prewitt operator for dataset preprocessing. order realize local overheated regions, introduces application DeepLabv3+ neural network algorithm based MobileNetV2 as backbone segmentation task first time. results show that achieves 65.36% average crosslinking rate (mIoU), 78.75% accuracy, 0.78 s per-iteration processing time, 22.4 MB parameter size. provides substantial advantages terms reduced computing memory resources detection Ultimately, method proposed enables EAST; identifies areas prevents damage or other critical components due ensuring

Language: Английский

An intelligent federated learning boosted cyberattack detection system for Denial-Of-Wallet attack using advanced heuristic search with multimodal approaches

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 24, 2025

Language: Английский