The point of this study is to improve quality control of rubber hoses in a chosen company. The objective is to maintain strict wall thickness constant throughout the circumference and in conformity within the norm. Currently, measurements are taken by staff using calipers. The work will be preceded by analysis of that means of quality control in the chosen company. At the end, QC shall be achieved through computer vision systems, with as little human effort as possible, and in as short amount of time as possible. To start, the 3D model of a new stand will be created. The requirements for it is to be ergonomic, easy to use, relatively easy to build, reliable and easy to maintain (e.g. the glass that supports the hose during measurement must be nonscratchable and easy to clean). Stand will include HMI (Human Machine Interface) that will show all important parameters of measured hose in accessible form. Image processing and analysis will be done using an algorithm created in Adaptive Vision Studio program, with the use of Deep Learning add-on. At the end, the results of modernisation will be validated.
