Project Splinescan

QUICKLINKS:

A Brief Introduction

In June 2002 a sun-starved computer engineer walked into a small room armed with an old PC, a 20 year-old video camera, some lasers, glass rods and an awful lot of duct tape. His mission was to create a system capable of converting a real world object into a computer model that could be manipulated, stored and eventually reproduced using stereolithography.

After several hours of gluing, taping and generally walloping things with a hammer – Project Splinescan was born. Originally running on a Microsoft Windows machine (K6-500), the prototype scanner proved the concept of low cost 3D scanning with a few rough scans. The challenge now is to convert the whole project to linux and produce some maintainable code that can be open-sourced and documented properly. Also on the agenda are methods to texturise and surface the scanned models.

The original prototype scanner used a class 3b surveying laser, and a free wheeling turntable. The prototype was principally used to develop the scanning and encoding algorithms, and prove the plausibility of the project through disposable prototyping.The new scanner will be able to capture colour scans with a much higher level of accuracy. The turntable on the scanner can also be replaced with a motion controlled version, which makes interpolated scanning possible. Hole filling, smoothing and multiple output formats are on the current list of things to do.

SNIPPETS:

GOING TO PYTHON
The splinescan project is being redesigned from the ground up using PYTHON. I tried experimenting with C but in the end, the programming started to overtake the project, and was becoming inaccessible. Python is an interpreted language, runs on most platforms, is easy to learn and is ideally suited to the scanner project.

V4L
The capture system uses streamer to gather data - While V4L programming was ok to do, there is very little need to reinvent the wheel and streamer does exactly what the project needs it to do.