16/11/2022 – „Eddyeyes“ – vision and artificial intelligence
AI to classify defects automatically
Gone are the days when defects were detected with eddy current coils and the results had to be interpreted with varying degrees of expertise by the inspector. This does not mean that eddy currents are not valid; on the contrary, they are still the main inspection technology for hot-rolled long products. No other is as reliable, as economical and as widespread.
But the times are also gone when images of defects could be obtained either by machine vision or by photographing those detected by eddy currents. A coil of more than thirteen kilometers could have hundreds, thousands. of defects, or potential defects, serious, light, false positives, etc. It is not productive to dedicate human resources to look at and interpret thousands of images that are not always of high quality.
Today, the key to quality control of this type of product is the use of artificial intelligence (AI) to classify defects automatically with high percentages of accuracy. This information is crucial to deliver the right product for each customer, reduce complaints and improve market loyalty.
Isend has been developing its “Eddyeyes” systems since 2014 under these premises: eddy currents as primary detection system, artificial vision as support system and artificial intelligence with advanced machine learning algorithms for defect classification in real time. Additionally, this information can be shared between customers and suppliers, taking quality control and traceability to a higher level.
With “Eddyeyes” a rolling mill can save millions dollars a year, reduce the CO2 emissions and become more competitive against low quality competitors.