Development of a Vision-based Monitoring System for Quality Assessment of 3D Printing

Abstract

Additive Manufacturing (AM), also known as 3D printing, is a process of manufacturing parts and components by adding successive layers of material on top of each other until the final shape is achieved. The research target of this project is Fused Filament Fabrication (FFF), which is a specific type of Additive Manufacturing technology. FFF uses a filament of thermoplastic material, which is melted and extruded, then deposited layer by layer to create a 3D object. However, FFF has some limitations that need to be considered. For instance, the printing process can be time-consuming, and errors such as misalignment and incorrect slicing can occur, leading to complete failure and wasted time and material.

This thesis presents a vision-based monitoring system for FFF 3D printing quality assessment. The proposed system includes a simulation tool that generates simulated images of printed layers, along with feature extraction methods for assessing the size, shape and infill density of printed objects. The proposed system utilizes background subtraction for isolating the printed object from the background and estimating its size through pixel length analysis and bounding box calculation. The shape analysis of the printed objects is performed using the Fourier-Mellin transform (FMT) method. Moreover, the infill density is computed by combining foreground extraction and image thresholding methods, utilizing both camera and simulated images. The proposed system is able to analyse and examine the quality of 3D printing during the printing process and identify the defective printed object when deviates of 5 percent is detected in terms of the size, shape, and density of the printed object, alert the user to terminate the entire process and save time and cost. This new monitoring system provides an effective solution to improve the quality and efficiency of FFF 3D printing.