Title :
Computer image analysis for identification of wear particles
Author :
Thomas, A.D.H. ; Davies, Trevor ; Luxmoore, A.R.
Author_Institution :
Sch. of Eng., Univ. Coll., Swansea, UK
Abstract :
Wear debris in oil can be visually inspected by an expert to indicate the condition of machinery, reducing costs and preventing failure. Of particular interest are metal particles generated by wear. These particles fall into several well-defined classes, each class representing different wear conditions. The authors outline how the identification and analysis of these particles can be effectively performed by an image processing computer with the ability to make a `human-like´ diagnosis. An automated system could release an expert from this task, and produce quantitative data not revealed by the human eye. The automatic process involves three stages: digitisation of the field of view: identification of individual particles; and particle analysis using previously established rules, usually based on particle size and shape
Keywords :
computerised picture processing; maintenance engineering; computerised picture processing; diagnosis; digitisation; failure; image processing computer; maintenance engineering; metal particles; particle analysis; wear particles;
Conference_Titel :
Pattern Recognition for Binary Images, IEE Colloquium on
Conference_Location :
London