Development of visual inspection systems for 3D printing

Every day the popularity of 3D printing is growing. It is used in almost all fields of the modern science and techniques. Applications of this technology is widespread: from architecture and design, to Biomedicine and aerospace. Despite its popularity, the process of additive 3D printing usually takes a while. During this process we need to be sure that the printed object stays static, besides, we need to control the absence of contamination on the print head and print defects. In this paper we propose a solution to the problem of detecting collisions in the additive process of 3D printing through the development of a universal system of visual control. Technical process control is carried out by the introduction of visual feedback. We have described two methods, which are used in computer vision for recognition: it is the machine learning and tracking based on singular points. The joint use of these two techniques provides improved quality of collision detection. During this research we have developed two types of the labels — sealed labels and superficial labels. The relevance of this development is confirmed by the possibility of obtaining economic benefits from its integration, by reducing the amount of consumables in case of collisions. It is especially true in industries that use expensive printed materials.