Process Control-Based Embedding and Computer Vision-Based Retrieval of 2D Codes in Fused Deposition Modeling

Abstract

Abstract Information lithography in manufactured parts is a broad set of techniques for encoding sequences of bits as physical or behavioral features in physical parts. It is gaining increasing attention within the secure manufacturing community. In our prior work, we have shown that information can be embedded on the surface of additively manufactured parts by controlling the manufacturing process parameters. In this paper, we extend that work to show that the information embedded by controlling the speed of the printer head in the fused deposition modeling (FDM) process can be read using optical cameras. Specifically, we present a method for embedding and retrieving 2D codes, such as watermarks, on the surface of parts printed via the Fused Deposition Modeling process. Our approach is to embed information by locally varying printing speed to represent different bits, and retrieving the bits using simple optical cameras. We demonstrate it using QR codes printed using a low-cost commercial FDM printer and retrieve it using a consumer-grade DSLR camera and 2D paper scanner. The color image captured by the camera/scanner is converted into a QR code through two approaches: (a) morphological segmentation, and (b) training binary classifiers on experimental data. We observe that morphological segmentation is easy to implement, and results in fast classification, but is demanding in terms of having consistent imaging conditions (e.g., lighting) and tuning for each specific instance. On the other hand, the binary classification approach requires training data, but results in higher overall accuracy (∼92%). The results show that the proposed approach is an effective and inexpensive method for information lithography in AM parts, and it can be utilized to embed information such as serial numbers, watermarks, and any other meta-data on the physical parts.