This thesis is a two-part project that addresses the development of a vision system designed for use in quality inspection of assembled piston rings. The second part concerns concept development for feeding pistons inro a robot cell. Currently, inspection of assembled piston rings is carried out visually by operators, which is a monotonous and time-consuming task. The purpose of developing the vision system is to investigate whether it is possible to use machine vision and machine learning to detect defects in the pressed piston rings. The work on the vision system was carried out in several stages. It began with literature studies that progressed to data collection and training of models. The models where subsequently tested in order to derive lessons for the next training iteration. The purpose of concept development is to solve the problem of inserting pistons into a robotic cell with minimal manual intervention. Concept development began by defining a requirements specification that was used to generate concepts. Concepts that were not considered sufficiently good were then screened out. The remaining concepts were evaluated using a weighted concept scoring matrix. Both the vision system and the developed concepts are considered to have strong potential to become complete solutions, but further work is required Before implementation can be carried out.