Computer Assisted Orthopaedic Surgery (CAOS) systems improve the results and the standardization of surgical
interventions. Anatomical landmarks and bone surface detection is straightforward to either register the surgical space
with the pre-operative imaging space and to compute biomechanical parameters for prosthesis alignment.
Surface points acquisition increases the intervention invasiveness and can be influenced by the soft tissue layer
interposition (7-15mm localization errors).
This study is aimed at evaluating the accuracy of a custom-made A-mode ultrasound (US) system for non invasive
detection of anatomical landmarks and surfaces. A-mode solutions eliminate the necessity of US images segmentation,
offers real-time signal processing and requires less invasive equipment. The system consists in a single transducer US
probe optically tracked, a pulser/receiver and an FPGA-based board, which is responsible for logic control command
generation and for real-time signal processing and three custom-made board (signal acquisition, blanking and
We propose a new calibration method of the US system. The experimental validation was then performed measuring the
length of known-shape polymethylmethacrylate boxes filled with pure water and acquiring bone surface points on a
bovine bone phantom covered with soft-tissue mimicking materials. Measurement errors were computed through MR
and CT images acquisitions of the phantom. Points acquisition on bone surface with the US system demonstrated lower
errors (1.2mm) than standard pointer acquisition (4.2mm).