Ingeniería Civil Matemática certificada por 3 años hasta Diciembre de 2026
Sergio Nicolás Pincheira Tapia
Bachelor’s Thesis
Automated Classification of Coronographic Images and Statistical Analysis of Power Outages
Thesis Advisor
Lisandro Fermı́n
Héctor Olivero
Summary
Coronographic inspections are a relevant topic in the field of electricity, especially for assessing the condition of components in power distribution systems. The usual approach for conducting these inspections involves the use of specialized cameras to capture coronographic images, which are then analyzed and classified. However, evaluating the images individually consumes a significant amount of time.
Given this context, the present work is carried out in collaboration with Chilquinta and focuses on automating the classification of damages in coronographic images and analyzing a history of power outages through count process estimations in order to assess the potential impact of the inspections.
The project development is divided into two stages. The first involves the automation of coronographic image classification, which includes preprocessing, the application of Optical Character Recognition (OCR), and the implementation of an automated classification system integrated into a web platform. The second stage consists of analyzing power outages and performing estimations based on homogeneous and non-homogeneous Poisson processes to evaluate the potential impact of coronographic inspections.
Each development stage presents its corresponding results. It is worth noting that the automated classification system achieves excellent results on test images when the appropriate OCR is used, allowing for accurate classification of all available coronographic images. Additionally, the analysis through count process estimations provides relevant insights into the intensity of power outages across different zones and sectors, as well as the potential impact of coronographic inspections.
