A Custom-Made Photogrammetry Scanner to Support Paediatric Surgery

Abstract

Photogrammetry is a challenging technique used to digitalize objects and to build their surface points’ meshes by connecting digital photographs taken from different positions. The work presented here is the first phase of a collaboration between the OSI EECruces Hospital and the Faculty of Engineering in Bilbao to develop a portable scanner to perform digital scans of paediatric patients’ faces with enough accuracy to: (i) train surgery operations, (ii) analyse patients’ evolution and (iii) perform virtual reality surgery guides. This first phase deals with the construction and calibration of a custom-made scanner. The scanner was built around 12 Raspberry Pi cameras and single-board computers, assembled on different positions in a custom-made structure inside an ad hoc designed and 3D printed container. The Raspberries were connected and synchronized via Ethernet to allow the 12 photographs to be taken simultaneously. Additionally, a projector was used to generate a particular light pattern to aid connecting the different photographs and generating the resulting two 3D points’ meshes: (i) a real colour and texture of the test subject’s face mesh and (ii) a mathematical mesh. Calibration was performed with a rule placed onto the test subjects’ face. This first study was designed to test the accuracy of the photogrammetry scanner by comparing its performance to that of an industrial portable scanner. The promising results obtained indicate that the described scanner has a big potential to perform clinical studies.