A new automated system developed at Purdue University can detect cracks in nuclear power plants more accurately than other systems. Called CRAQ, the new system uses an advanced algorithm and machine learning to analyze remotely recorded inspection videos.