Visual Inspection is routinely carried out across industry to determine whether a structure, product, component or process meets the specified requirements. Typical examples include the detection of product defects in-service or during maintenance and as point of manufacturing in-process monitoring. Such inspection is usually carried out by a trained individual who has sufficient knowledge and experience to visually identify faults and non-conformant quality and performance.
There are a number of challenges associated with visual inspection such as accessibility constraints (height, confined space, hazardous environment etc.), increased susceptibility to human error (in the case of repetitive high frequency visual inspections) and recurring labour costs. Moreover, in industry the visual inspection activity is classified as essential but a non-value add step in the process.
Artificial Intelligence (AI) provides an opportunity to introduce innovation and new technology to the visual inspection process, offering a solution to challenges and requirements. The Centre for Modelling & Simulation (CFMS) has produced a demonstrator that uses a combination of computer vision and AI technologies to automate the manual inspection process. Captured by a camera mounted on a drone or crawling robot, the AI modules are fed with a series of images as an input, and processed with Convolutional Neural Networks (CNN) which are trained to learn the “rules for feature identification automatically” (rather than explicitly programmed rules) and identify objects (e.g. bolts, brackets etc.) and/or defects (e.g. corrosion, damage etc.) irrespective of their position or orientation in a real world inspection scenario.
Automated Inspection technology offers a range of benefits, and provides a more efficient and streamlined solution for inspecting difficult to access or hazardous environment and areas. It reduces both the inspection time, and the cost of inspection while increasing output and productivity, enabling the standardisation and automation of documenting assets, e.g. for condition of supply documentation. Automated Inspection technology is applicable across sectors, and will significantly benefit high value manufacturing, maintenance, repair and operations, in aerospace, automotive, construction and energy sectors, where visual inspection is highly utilised.