Colour digital image correlation method for monitoring of cultural heritage objects with natural texture

Abstract

Digital image correlation (DIC) is a well-established technique for the measurement and monitoring of displacements and strains in engineering objects. In technical applications, an object's surface is modified by applying random, artificial structure, and monochrome cameras are typically used. The challenge in the measurement of Cultural Heritage objects (CHO) lies in the necessity of using the natural surface texture (no modifications are allowed), which in most of the cases is far from the ideal random pattern required for the successful implementation of DIC. On the other hand, spectrally separated content of natural CHO textures may provide additional spatial information and support DIC analysis. In this work, we propose a colour image pre-processing path that provides the input data best fitted to high accuracy 2D DIC and 3D DIC analysis. For this task we employ cameras with a colour filter array (CFA) with a Bayer pattern. This pattern has twice the number of green photosites as red or blue arranged in a 2x2pixel repeating-unit. From this sparse colour sampling, a full colour image is produced by interpolating the unknown values in each color channel. The interpolation is achieved via any one of the many demosaicing algorithms available to calculate a red, green, and blue value for every pixel position that only data from one channel was captured. These three R, G, B channels can then be combined to create a monochrome image suitable for conventional DIC analysis. This work's main objective is to select the best demosaicing algorithms for the spectral and monochromatic channels to minimize displacement and strain reconstruction errors (STD and P/V). The systematic analysis will be performed for objects (mock-ups) with artificial texture and fully controlled displacements. The expected errors due to interpolation procedures will be identified. Next, the best demosaicing procedure will be implemented into the full processing path of colour DIC used for selected CHO with natural texture (parchments, oil paintings). The analysis results based on monochrome, spectral, and mixed spectral data channels will be compared from the point of view of their applicability to different types of natural texture and achieved measurement accuracy.