Robust Surface Detection for Variance Comparison and Dimensional Measurement

Abstract

This paper describes a robust method for creating surface models from volume datasets with distorted density values due to artefacts and noise. Application scenario for the presented work is variance comparison and dimensional measurement of homogeneous industrial components in industrial high resolution 3D computed tomography (3D-CT). We propose a pipeline which uses common 3D image processing filters for pre-processing and segmentation of 3D-CT datasets in order to create the surface model. In particular, a pre-filtering step reduces noise and artefacts without blurring edges in the dataset. A watershed filter is applied on the gradient information of the smoothed data to create a binary dataset. Finally the surface model is constructed, using constrained elastic-surface nets to generate a smooth but feature preserving mesh of a binary volume. The major contribution of this paper is the development of the specific processing pipeline for homogeneous industrial components to handle large resolution data of industrial CT scanners. The pipeline is crucial for the following visual inspection of deviations.