Spherea US (formerly Konrad Technologies USA), with ADS, G2CPU and Radx Technologies, has developed an AI-based computer vision approach for testing EO sensors and cameras using NI (Emerson) hardware. The system enables automated, real-time anomaly detection that significantly improves coverage, repeatability and reproducibility while reducing test costs in both engineering and production settings, even with multiple devices per station. Unlike conventional AI models requiring thousands of training images, this solution detects defects after only seconds of in-line observation of known-good units. This makes it highly efficient and well suited for scalable production line applications.

• How AI-based vision improves EO sensor defect detection with greater coverage, repeatability and reproducibility
• How PXIe hardware and GPUs enable automated, high-throughput testing of multiple high-data-rate devices
• How minimal training data allows fast, scalable defect detection in production environments