digital fabrication.pdf

UNIT-1: Introduction to Additive Manufacturing and CAD  Lesson 1.4 Reverse Engineering for 3D Printing 1.4.1 INTRODUCTION Reverse engineering (RE) is a process that involves creating a precise and accurate CAD (Computer-Aided De...

UNIT-1: Introduction to Additive Manufacturing and CAD  Lesson 1.4 Reverse Engineering for 3D Printing 1.4.1 INTRODUCTION Reverse engineering (RE) is a process that involves creating a precise and accurate CAD (Computer-Aided Design) model from an existing physical object or part. This technique is often employed for various purposes, such as modifying a design, understanding an object's architecture, or analyzing its functionality. With the advent of 3D printing, replicating and fabricating copies of parts has become faster and more efficient than ever before. The initial step in reverse engineering involves acquiring the geometric data of the physical part to develop its CAD model. For simple objects composed of primitive geometry, this data can be directly obtained using measuring tools. However, for more complex parts, 3D scanning methods are employed. In the medical field, geometric data of human body parts is gathered from CT or MRI scans. Reverse engineering has found wide-ranging applications across various domains, including rapid product development, casting, CNC machining, medical imaging and demonstration, medical tooling, and part inspection, among others. This unit aims to provide an understanding of the principles, techniques, and applications of reverse engineering in today's fast-paced technological landscape. 1.4.2 REVERSE ENGINEERING APPROACHES It is evident that when creating a CAD model from existing parts composed of primitive geometry, one can simply measure the geometric details using various measurement tools, and then create a solid model using any CAD software. However, for parts with complex geometry, such as freeform surfaces or natural shapes that are difficult to measure, 3D scanning or other methods can be employed to acquire the necessary geometric data. Point cloud data from a part's surface may be gathered using a process called 3D scanning. It is frequently necessary to perform preprocessing operations on this point data, such as outlier and spike removal, point data reduction, and organization. When using a mesh-based surfacing technique, the well-organized point data is meshed, or it is segmented for a curve-based surface modeling approach. This creates a CAD model. Fortunately, owing to sophisticated Page 1/7 computational techniques and algorithms, pretreatment is virtually no longer necessary with the point cloud processing technologies that are now accessible. Figure 1 Reverse engineering steps 1. Utah teapot. (2024, February 14). In Wikipedia. https://en.wikipedia.org/wiki/Utah_teapot, 2. https://inspirationtuts.com/the-important-utah-teapot/ Following the scanning process, the point cloud data is converted into a surface model or voxel model of the geometry. This surface model is then transformed into an STL file or other 3D printing-friendly formats, before being 3D printed. Contact and non-contact devices are the two types of devices used to obtain surface data from a component. Alternatively, point cloud data for layered or concealed objects—like human bones—that can't be scanned with 3D scanners is obtained using medical imaging techniques. a) Coordinate measurement machines (CMMs) are commonly used for inspection and tolerance purposes. These NC-driven devices can gather point data with many microns of precision. However, they are intrinsically sluggish at gathering point data since each Page 2/7  sampled point requires physical contact with the part's surface. As a result, CMMs are inefficient in measuring freeform components and are primarily employed for automated inspection of simple fundamental forms such as slots, steps, holes, and pockets. b) Non-contact devices, on the other hand, do not require physical contact with a part in order to acquire point data, making them a more realistic alternative for rapidly gathering huge volumes of data. Laser scanners are the most used non-contact measurement equipment because they can quickly and accurately collect large volumes of point data. Laser scanning equipment' accuracy has improved, and their use in industry has increased dramatically. c) Compressed tomography (CT) scanners are used to scan the material composition and interior geometry of components. The human body may be seen using medical CT data, which can also be utilized to design surgical implants. High-power industrial CT scanners may be used to securely get information on the interior shape, porosity, and density of industrial items. Reverse engineering in the medical field is a specialized process that differs from traditional reverse engineering techniques. It involves creating accurate 3D models of body parts or internal structures based on CT or MRI scans, which are typically available in Digital Imaging and Communications in Medicine (DICOM) format. The process is essential for various applications, such as surgical planning, implant design, or education. Figure 2 CT scan to 3D model conversion steps The first stage of reverse engineering for medical parts involves obtaining and processing CT or MRI scans containing the required anatomical details. An experienced operator Page 3/7  segments these scans based on pixel values, which correspond to the x-ray attenuation coefficient of the body part. This segmentation process is crucial as it isolates the desired area of the CT scan file for further processing. Once the segmentation is complete, the operator creates a mask representing the segmented part of the CT scan images' 3D array. This mask is essential for generating the point cloud data, which is then converted into a voxel model. Voxel models represent 3D objects on a computer screen by dividing the object into a three-dimensional grid, with each cell in the grid called a voxel. This representation helps visualize the solid geometry of the scanned body part. After creating the voxel model, the point cloud data is converted into an STL (Stereolithography) file format. STL files are commonly used in 3D printing and additive manufacturing processes, as they describe the surface geometry of a 3D object without representing colour, texture, or other attributes. Finally, the STL file is used for 3D printing the physical model of the medical part. This 3D printed model can serve various purposes, including assisting surgeons in planning procedures, designing custom implants, or educating patients about their anatomy or upcoming surgeries. Reverse engineering of medical parts is a complex and essential process that requires specialized software, expert knowledge, and meticulous attention to detail. It involves acquiring and processing medical data, segmenting images, creating a mask, converting point cloud data into a voxel model, generating an STL file, and 3D printing the final model. The resulting 3D models can significantly enhance medical professionals' understanding of anatomical structures and improve patient care. 1.4.3 REVERSE ENGINEERING OPERATIONS FOR CLOUD DATA The point cloud and mesh models are representations of the geometry in terms of 3d space data points. Each point in the point cloud has its coordinates about the principal axis. Sometimes it is required to process the acquired data points by applying some operations. Here we are discussing the general operations which usually employed over point cloud data. 1. Noise reduction: Point cloud data obtained from scanning devices or environmental factors can often contain unwanted noise. To improve the accuracy of this data, noise reduction algorithms such as statistical outlier removal and moving least squares are used to filter out the noise. Page 4/7 2. Registration: When creating a 3D scan of a large or complex object, it may be necessary to take multiple scans in order to capture the entire geometry. To form a complete and seamless point cloud, these individual scans are aligned and merged using registration algorithms like Iterative Closest Point (ICP) or Global Registration. 3. Down-sampling: Point cloud data with high resolution can be quite large and require significant computational resources to process. To reduce the number of points in the point cloud while retaining its overall shape and features, down-sampling algorithms like voxel grid filtering and random sampling are employed. 4. Normal estimation: Surface normal are vectors that are perpendicular to the surface at every point. They are important for different processing tasks, such as mesh generation and surface reconstruction. Normal estimation algorithms, like Principal Component Analysis (PCA) and normal estimation using k-nearest neighbours, compute the normal vectors for each point in the point cloud. 5. Feature extraction: Point clouds are 3D data sets obtained through LiDAR or photogrammetry. Identifying features in point clouds like edges, corners, and planes is useful for object recognition, segmentation, and registration. Feature extraction algorithms like FPFH and SIFT detect and describe these features. FPFH is effective in detecting local geometric features, while SIFT is useful in detecting scale-invariant features. These algorithms are crucial in 3D modeling, autonomous driving, and robotics. 6. Segmentation: At times, the point cloud may include various objects or parts that require segregation for further processing. Segmentation algorithms, such as region growing, Euclidean clustering, and RANSAC (Random Sample Consensus), are employed for partitioning the point cloud into significant segments based on geometric criteria or feature similarity. 7. Surface reconstruction: When converting point cloud data into a mesh model, it's necessary to create a continuous surface that accurately represents the geometry of the object. This process is called surface reconstruction and it involves the use of specialized algorithms that estimate the surface by connecting neighboring points and forming a mesh of connected triangles. There are several surface reconstruction algorithms available, including Poisson Surface Reconstruction, Ball-Pivoting Algorithm (BPA), and Marching Cubes. These algorithms are crucial in creating accurate and detailed 3D models of objects from point cloud data. To ensure that your 3D printing projects turn out accurate and of high quality, it's important to have a solid understanding of basic operations and algorithms for processing and manipulating point cloud data. Luckily, there are several software tools and libraries available to help facilitate these tasks, including Point Cloud Library (PCL), MeshLab, and CloudCompare. These Page 5/7 powerful tools offer a wide range of point cloud processing functions and algorithms, making it easier than ever to create precise and detailed mesh models for your 3D printing needs. In conclusion, reverse engineering for 3D printing is a powerful technique that has revolutionized various industries, including product development, manufacturing, and medical applications. By acquiring precise geometric data through advanced 3D scanning methods, such as laser scanners and medical imaging techniques, it is now possible to create accurate and high-quality CAD models of complex objects and internal structures. The subsequent conversion of this data into surface or voxel models and 3D printable formats has enabled rapid prototyping, customization, and the development of innovative solutions across a broad range of domains. As technology continues to advance, the potential applications and benefits of reverse engineering for 3D printing will undoubtedly expand, contributing to a more efficient, personalized, and cutting-edge future in design and manufacturing. Assignment 1.4.1: - Depending on the complexity of the object and the resources available, describe the method you would use to obtain the geometric data of the object, such as manual measurement or 3D scanning. Explain your choice and the potential challenges or limitations of the chosen method. - Choose a household object with moderate complexity (e.g., a kitchen utensil, a small tool, or a decorative item) that you believe would benefit from reverse engineering and 3D printing. Describe the object and explain why! you selected it for this assignment. Discussion Topic 1.1 Discuss the use of 3D printing processes in Medical domain. References Additive Manufacturing Technologies by I. Gibson, D. Rosen and B. Stucker (Springer). Additive Manufacturing of Metals: The Technology, Materials, Design and Production by L. Yang, K. Hsu, B. Baughman, D. Godfrey, F. Medina, M. Menon and S. Wiener (Springer). 3D Printing Technology, Applications and Solution by Rafiq Noorani (CRC Press). Page 6/7