Scientific AI — Deep Capability
Electroluminescence
Inspection System.
Solar panel defects are not always visible. Electroluminescence imaging
captures the electrical activity of each cell under forward bias, making
micro-cracks, finger interruptions, and material degradation visible as
distinct luminescence patterns.
AKIHARU's EL inspection system applies convolutional neural networks trained
on domain-specific EL image datasets to classify defect types with high
spatial precision — enabling systematic quality assessment at the
manufacturing and field deployment stages.
EL image acquisition and preprocessing
CNN-based defect classification
Micro-crack and finger interruption detection
Physics-informed feature extraction
Batch panel inspection pipeline
Structured inspection reports