Lei Xia

In recent years, few-shot semantic segmentation of strip steel surface defects (S³D) has posed significant challenges, primarily due to the wide diversity of scenarios and the inherent complexity of defects. However, existing semantic segmentation methods have not effectively addressed these issues. In this paper, we propose a few-shot segmentation method called the Multiscale Adaptive Prototype Transformer Network (MAPTNet), which aims to integrate multi-scale feature aggregation and enhance the adaptability of defect detection across diverse and complex defect scenarios. Specifically, we introduce an Adaptive Prototype Transformer (APT) module to locate defects better and generate more refined feature representations. This module effectively explores the relationships between support and query sets and transforms them into optimal prototypes. Additionally, through hierarchical feature fusion to capture more detailed and contextually relevant S³D information, our MAPTNet improves the discriminability of defects at multiple feature scales. To further strengthen the learning process across different feature layers, our MAPTNet incorporates a deep dual supervision mechanism, facilitating effective optimization at both intermediate and final stages. Extensive experiments on the FSSD-12, Surface Defects-4i, and ESDIs-SOD datasets demonstrate that our network achieves state-of-the-art performance, significantly outperforming existing approaches. The code is available at https://github.com/hhhjjc/MAPTNet.

Accurate online junction temperature detection of high-power Insulated Gate Bipolar Transistor (IGBT) modules is necessary for the health management system of key equipment. The voltage rise time (t _rv ) is a good temperature sensitive electrical parameter (TSEP), which is of great significance for its accurate online extraction. In this manuscript, a novel IGBT junction temperature detection based on the high-frequency (HF) model of inductor element is proposed. Firstly, the equivalent model of the inductor element in the application circuit at high frequency is demonstrated. Secondly, based on the HF model of the inductor element, it is proposed and verified that the t _rv in the IGBT turn-off process is consistent with the high frequency current pulse signal on the inductor element at HF, that is, the t _rv can be indirectly obtained by extracting the high frequency current pulse signal. Finally, the feasibility for voltage rise time extraction and junction temperature detection through high-frequency isolation coil is verified by experiments and simulations. The research work in this manuscript is helpful to the on-line junction temperature detection, and has certain theoretical significance and practical engineering value.