Classification of Defective PV Cells in EL Imagery

Classifies solar cell images based on their defectiveness levels.

This project applies computer vision and machine learning to the ELPV dataset to classify solar cells by defect severity. The objective is to predict PV cell health from electroluminescence images, enabling early detection of defects that impact performance and reliability. By evaluating multiple modeling approaches, from general-purpose classifiers to deep convolutional networks, the work benchmarks accuracy, robustness, and real‑world applicability for automated PV module inspection.

⬇️ Contributions

Role: Associate Team Lead

I led one of two pipelines: Deep CNN classification, and contributed extensively to the thesis report.

⬇️ GitHub Page

For full project details, dataset, findings, and methodologies, see the GitHub page.

GitHub - denniemok/cnn-classifier-elpv: This project explores various machine learning and computer vision techniques, such as classifiers and CNNs, to classify defective solar cells in EL imagery using the ELPV dataset.GitHub

⬇️ Thesis Report