Fully Automatic Detection of Anomalies on Wheels Surface using an Adaptive Accurate Model and Hypothesis Testing Theory
Proc of European SIgnal Processing Conference (Eusipco)
IEEE / Eurasip
This paper studies the detection of anomalies, or defects, on wheels' surface. The wheel surface is inspected using an imaging system, placed over the conveyor belt. Due to the nature of the wheels, the different elements are analyzed separately. Because many different types of wheels can be manufactured, it is proposed to detect any anomaly using a general and original adaptive linear parametric model. The adaptivity of the proposed model allows us to describe accurately the inspected wheel surface. In addition, the use of a linear parametric model allows the application of hypothesis testing theory to design a test whose statistical performances are analytically known. Numerical results show the accuracy and the relevance of the proposed methodology.