Crash Course: 3D Scanning with the HDI LMI Advance


As the newest member of the 3D Proven Systems team I've been getting familiar with 3D Scanning, working on projects for our upcoming Proto Vault --more on what that is another time. Many assume that 3D Scanning is as simple as placing an item on a platform, pressing a button, and watching as a perfect 3D model appears on the computer screen. Scanning has come a long way in being user friendly, but it's still not that simple. I can tell you from personal experience it's as much a hands on craft as it is an automated technology. Here's what I learned from my crash course with the HDI LMI Advance: 

Precision Scanners Require Calibration: The HDI can capture the ridges and valleys of a fingerprint or a building, but it has to be calibrated based on the size of the object being scanned. Camera position can be adjusted to increase or decrease the focal point, and lenses can be changed to different scales of zoom. During Calibration, the scanner and software work together as a technician "grids-in" and focus a field of view appropriate for the object to be scanned so the scanner can accurately measure the distance between it and the object being scanned. This can be a lengthy process


Scanning Creates a 3D Model One Shot at a Time:  Our scanners project white light on an object to take individual 3D "pictures" from different angles. The 2 focused cameras use the white light grid to translate information on how far the light has traveled into 3D data. Capturing scans of an entire object means adjusting its position and angle(s). 

Scans Require Compositing to Create a Useful Model:  Scans are immediately imported into FelxScan3D, the software running the HDI scanner, which is used to combine the individual 3D "pictures." If the program recognizes the geometry, it will align the most recent scan with existing scans. Complex geometry or geometry without overlapping data will result in failure to fit the "puzzle" of scans; therefore, it must be manually positioned as close to accurately as possible before attempting to run an alignment. If the new scan still does not snap to alignment, it most likely means more scans are needed of the area to define that geometry. An effective method to help align scans is working from easily definable surfaces, and moving out to the rest of the object from there. After every scan is properly aligned, groups are combined to form a complete mesh model.   

You're Still Not Done: Once scans are combined, most of an object's geometry will be defined; however, it it extremely common to have tiny holes in the mesh model due to unavoidable factors like deep undercuts or internal features. The 3D scanner will only pck up data taht it can "see" directly. These can be filled either by trying to take additional scans or by using Flexscan 3D's hole filling tool. Using this tool offers both the option to automatically or manually fill holes.   


Finally: When a scan is finalized it can be exported in a variety of file formats including .obj and .stl. Files from a 3D scan can be sent straight to a slicing program and then 3D printed. The scans can also be imported into CAD programs like Solidworks and digital sculpting program like Z Brush, or even Oculus Medium. I was truly amazed learning how a 3D scanner can turn an object into a digital model which can be stored, manipulated, measured, printed out, scaled, reverse engineered, or manipulated so fluidly. In expert hands, there is nothing which can't be done.