| ABSTRACT |
INSIGHT: INSULATION INTELLIGENCE FOR GROWTH AND HIERARCHICAL TREE-TRACKING
A Cable Insulation Health Inspection System and method are disclosed for the automated detection, tracking, and quantitative analysis of electrical treeing in high-voltage cable insulation, such as cross-linked polyethylene (XLPE) and low-density polyethylene (LDPE). To address technical challenges related to image noise, optical occlusions, and limited datasets, the system integrates an artificial intelligence module utilizing Conditional Generative Adversarial Networks (cGANs) for high-fidelity image augmentation and synthesis. A core hierarchical segmentation engine employing a Cascaded Multi-scale Line Tracking Algorithm (c-MSLTA) is utilized to accurately isolate complex treeing structures from noisy backgrounds across multiple resolutions. The system provides automated measurements of critical degradation parameters, including tree length, width, and area, and facilitates temporal growth-rate analysis to predict the remaining useful life of the insulation. By incorporating an Image Quality Assessment (IQA) framework for performance verification, the invention offers a repeatable and precise diagnostic tool for early-stage insulation failure detection in power system infrastructure.
|