Rochester Institute of Technology
Defect Detection of Photovoltaic Panels by Image Processing
Abstract
dc:description.abstractA key component of the transition towards cleaner and more sustainable power sources, driven by the global demand for such energy, has been the fast improvement in the installation of solar photovoltaic (PV) plants. Ensuring the dependability of photovoltaic panels becomes crucially important when solar installations develop in scale. Photovoltaic panels have flaws that reduce their effectiveness and shorten their lifespan, but they are durable and long-lasting otherwise. Solar plants need to work as efficiently as possible with low downtime and to want solar energy to be viable in the long run, the issues shall be fined a fixed promptly. Quality monitoring has always relied on human inspectors. The speed, accuracy, and scalability issues that plague human inspection methods become more apparent as solar projects expand in size. The goal of this research is to improve solar panel flaw identification using cutting-edge image processing techniques. Getting beyond these problems is its aim. The project used Convolutional neural network (CNN). We will use ResNet50. ResNet50 is a convolutional neural network (CNN). It is a specific architecture within the CNN family that is designed for tasks related to computer vision, such as image classification, object detection, and image segmentation. The datasets that will be used in these sections are collected from Kaggle and from DEWA Solar PV Plants. The data will be subjected to Data Preprocessing, Data Cleaning and Organization, Data Augmentation and Feature Extraction. further building the model and go through the training, validation and testing process to ensure a high-performance model that accurately predicts the defects.
Degree
thesis:*- Name thesis:degree_name
- Professional Studies (MS)
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Graduate Programs & Research
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- AlMarzooqi, Hamad
- Contributors dc:contributor
-
- Ehsan Warriach
Subjects
dc:subject × 4Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.rit.edu/theses/11773