Automated Inspection of Photovoltaic cells using Deep Learning Algorithm
Bibliographic record
Abstract
The increasing global demand for renewable energy has increased the need for dependable and efficient inspection of photovoltaic (PV) cells. Solar deployment at scale requires automation, as manual inspection is labour-exhaustive, slow, and has the risk for human error. In this paper, we present an automated inspection framework for PV cells based on deep learning algorithms with the goal of reducing human involvement and to increase the efficiency of the inspection process. Our framework is built on the inspections performed in the base paper for the PV cells and even with the lesser computational cost and account for comparable accuracy when assessing the inspection characteristics of the PV cells, while minimizing the computational complexity of the algorithm. To ensure reliable defect classification, we utilize EfficientNet-based models with extensive data augmentation techniques and optimal training methods. Our process achieves high classification accuracy with decreased computational costs, which enables scalability and cost-effectiveness in PV cell inspection, as shown by experiments against the ELPV dataset. Additionally, using electroluminescence (EL) imaging creates a perfect identification of manufacturing defects in PV cells to detect microcracks, broken regions and inactive areas that cannot be seen with the naked eye. This reduces false negatives and reliably identifies defective cells. The lightweight aspect of the proposed model provides practicality for integration in production lines to help manufacturers automate quality control processes, reduce labor expense and visual inspections and ensure systemic product reliability.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".