Using Satellite Imagery to Improve Local Pollution Models for High-Voltage Transmission Lines and Insulators

This paper addresses the regression modeling of local environmental pollution levels for electric power industry needs, which is fundamental proper design and maintenance high-voltage transmission lines insulators in order to prevent various hazards, such as accidental flashovers due resultant outages. The primary goal our study was increase precision models this application area by exploiting additional input attributes extracted from satellite imagery adjusting methodology. Given that thousands different can be images, only a few are likely contain useful information, we also explored suitable feature selection procedures. We show combination attribute methods (relief, FSRF-Test, forward selection), (random forest M5P trees), methodology (estimating field-measured values target variables rather than their upper bounds) significantly total accuracy, measured correlation between estimated true variables. Specifically, accuracies dramatically rose 0.12–0.23 0.40–0.64, while relative absolute errors were conversely reduced (e.g., 1.04 0.764 best model).