In a manufacturing process defects are inevitable. The process of detecting and minimizing these are challenging and expensive. Therefore, it would be useful if the defects could be detected automatically with a machine vision system in the production environment. The purpose of the report is to evaluate potential vision based solutions that can identify these surface defects on metal surfaces as well as choose technology for Morakniv’s production. Furthermore, suitable locations for the vision system will be recommended and finally the investment will be evaluated from the cost of the system itself compared to the savings the system would generate.The report is a case study in which abduction is used to collect qualitative data from relevant literature, interviews, and observations to answer what technologies are available and to explore potential locations to install the system. Quantitative data from Morakniv and vision systems is used to evaluate the investment.The results from the study of the literature, interviews and an experiment show that vision with ai technology (machine learning) is best suited for Morakniv’s use case because of the complexity and variety of the defects. The locations most suitable for the system are presented from the interviews and observations and are further analyzed with a table in chapter 7. The investment is modelled with approximated quantitative data and later analyzed and motivated.Finally, recommendations for further work/study are laid out, the purpose of which is mostly to gain a deeper understanding of the training/installing process. The same work will also help in future dialogues with suppliers and experts of machine vision systems.