System and method for cyber-physical inspection and monitoring of nonmetallic structures

An inspection and monitoring system for analyzing a sequence of electromagnetic spectrum image frames of a nonmetallic asset and detecting or predicting an aberration in the asset, including a detected or predicted location of the aberration, the system comprising:

• a first machine learning system configured to receive the electromagnetic spectrum image frames and apply a convolutional neural network (CNN) to the electromagnetic spectrum image frames to stratify the electromagnetic spectrum image frames into abstraction levels according to an image topology and output first aberration determination information;
• a second machine learning system configured to apply one of a recurrent neural network (RNN) to the electromagnetic spectrum image frames to detect patterns in electromagnetic spectrum images over time and output second aberration determination information;
• an inspection and monitoring unit configured to generate an aberration assessment based on the first and second aberration determination information and transmit the aberration assessment to a communicating device;
• and
• a cyber-physical ambiance inspector that receives gas profile data indicative of a gas emitted from or by the asset, wherein the cyber-physical ambiance inspector is configured to predict the gas profile for the gas that might leak from or be emitted by the asset.

A non-transitory computer readable storage medium storing inspection and monitoring program instructions for causing a computing device to analyze a sequence of electromagnetic spectrum image frames of a nonmetallic asset and detect or predict an aberration in the asset, including a detected or predicted location of the aberration, the program instructions comprising the steps of:

• receiving the electromagnetic spectrum image frames by a pair of machine learning systems of different types;
• applying a convolutional neural network (CNN) to the electromagnetic spectrum image frames to stratify the electromagnetic spectrum images into abstraction levels according to an image topology and output first aberration determination information;
• applying a recurrent neural network (RNN) to the electromagnetic spectrum image frames to detect patterns in electromagnetic spectrum images over time and output second aberration determination information;
• generating an aberration assessment based on the first and second aberration determination information;
• transmitting the aberration assessment to a communicating device;
• receiving gas profile data indicative of a gas emitted from or by the asset using the electromagnetic spectrum image frames;
• and
• predicting the gas profile for the gas that might leak from or be emitted by the asset;
• wherein the aberration assessment includes prediction of an aberration and a location of the aberration in or on the nonmetallic asset.