Papers
7
Total Citations
117
H-Index
5
About
Hongxin Ji’s research sits at the intersection of intelligent robotics, non-destructive inspection, and deep learning, with a focused mission to enable autonomous robots to operate inside the most inaccessible industrial environments—large oil-immersed transformers. Ji’s major contributions center on developing miniature patrol robots capable of visually and ultrasonically inspecting internal transformer defects, such as insulation degradation and carbon-trace formation. By integrating advanced signal processing techniques—including empirical mode decomposition, wavelet denoising, and generalized cross-correlation algorithms—Ji has pioneered robust 3-D spatial localization methods that allow these micro-robots to navigate and position themselves accurately within metal-enclosed, oil-filled chambers. On the visual side, Ji has advanced deep-learning-based defect detection and image segmentation, notably through improved MSRCR enhancement and novel models like DCMC-UNet, which dynamically fuse features and adapt to challenging lighting conditions. With over 100 citations across a rapidly growing body of work—including the highly cited 2020 study on human posture recognition using fuzzy logic and machine learning—Ji’s research demonstrates both breadth and practical impact, offering transformative solutions for the predictive maintenance of critical power infrastructure.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7