Deep Learning Implementation for Snail Trails Detection in Photovoltaic Module

The degradation of fossil fuel reserves and the increasing enthusiasm for the development of renewable resources have led the world to seek and create renewable resources. Among the alternative energy sources available, solar energy is claimed to be the main choice that can replace fossil resources to face the demands of power consumption needs at this time due to the unlimited supply of solar energy. However, the awareness to check the condition of Photovoltaic modules is still low. Detecting the snail trails is necessary to determine the performance of snail trails as an initial step to prevent ongoing damage. This study aims to develop a snail trails detection system on photovoltaic module images using the machine learning method called as deep learning with the YOLO (You Only Look Once) algorithm version 3. There are several important processes required in the YOLOv3 method, which are image annotation, training data, and testing data. The result of detecting snail trail objects on the photovoltaic module obtained 99.7% of accuracy. This value suggests that this method can be proposed to contribute to the implementation of the development of effective damage detection research early on solar panel modules.