Reading Time: 2 minute

Large solar panel installations are essential for our future energy production without the large carbon dioxide emissions we are producing today. However, microscopic fractures, hot spots, and other surface defects can expand over time, leading to reduced output and even failures if not detected. Manivannan Sivan’s solution for solving this issue revolves around using computer vision and machine learning to detect small surface defects before automatically reporting the information.

Sivan compiled his dataset by first gathering images of solar panels with visible cracks using an Arduino Portenta H7 and Vision Shield and then drawing bounding boxes on each. From here, he trained a MobileNetV2 model with the addition of Edge Impulse’s new FOMO object detection algorithm for better performance. He was able to further improve the accuracy of the model by supplementing the data with images taken at different camera angles and lighting conditions to avoid mistaking the white boundary lines for cracks.

After testing and deploying the model from Edge Impulse Studio to his Portenta H7 board, it successfully detects cracks on the solar panel surface about 80% of the time. In the future, Sivan may add other features that take advantage of onboard connectivity to communicate with external services for faster response times. You can read more about the project here.

Website: LINK