---

Braking behavior as well as life cycle of a railway brake system depends on the quality of the used brake pads. These qualities are evaluated today in a lot of experiments and a classification (good / bad) is mainly done by the know-how of the railway engineers.This work presents an image processing approach for classification of railway brake pads. It is based on the processing of thermal images monitored during a brake application. The work can be divided into two parts: In the first part the performance of different classical differential filters is evaluated to show, which one could give significant information. An attempt has also been made for finding symmetrical information in thermal images using gradients. In the second part, different methods for brake pads classification are proposed, combining significant information calculated from different image processing. This proposed quality factors are examined and compared using the measured image-sequences of brake applications with bad and good brake pads. This work has shown some useful results and can be used for the classification of brake pads and the definition of quality factors.

---

Sarfraz Majeed is working as Lecturer and is a Phd candidate in Computer Science department at University of Management and Technology Lahore Pakistan. He received Master of Science(M.Sc) Information Technology degree from University of Applied Sciences Lemgo Germany in 2008.His area of interest includes Algorithms Performance,Distributed Systems.

---