The development of the world around us accelerates. To ensure security of new technologies, to introduce improvements of those already in existence and to eliminate undesirable events during construction phase and operation of new machines that aid our evolution, we have to devise new and faster methods of researching the processes. The combination of computer image analysis with experiences of a relatively new branch of science, yet much needed one, such as tribology, is justified. Current computerized vision systems, replace and support people in multiple processes of analysis and object class recognition in such branches of science as: crystallography, metallography, medicine, agricultural engineering and others. By using this method, the range of perception of our sight is expanded to notice differences that cannot be detected during direct observation, moreover, it also enables automatic and mass analysis, with classification of hundreds of thousands of images based on various criteria, without limitations that a human being is prone to, such as emotions or fatigue. The article presents a method of using computer image analysis to assess quantitative and qualitative wear of belt tensioner rollers.