This project explores the application of machine learning and deep learning techniques to identify the need for cleaning insulator chains in high-voltage towers.
The project includes the development of a support algorithm that utilizes neural networks to detect the insulator chains and identify the type of impurity on the disc surfaces.
The creation of a simulated database was necessary due to the scarcity of available real images. The results obtained demonstrated high accuracy in detecting insulator chains and identifying impurities, enhancing safety and efficiency in maintaining these essential components for the operation of the high-voltage power grid.
https://www.kaggle.com/datasets/hericlesfelipe/insulator-dataset-simulatedhttps://www.kaggle.com/datasets/hericlesfelipe/insulator-dirty-dataset-simulated