Accurate measurement of curvilinear shapes by Virtual Image Correlation

Abstract

Abstract. The proposed method allows the detection and the measurement, in the sense of metrology, of smooth elongated curvilinear shapes. Such measurements are required in many fields of physics, for example: mechanical engineering, biology or medicine (deflection of beams, fibers or filaments), fluid mechanics or chemistry (detection of fronts). Contrary to actual methods, the result is given in an analytical form of class C ∞ (and not a finite set of locations or pixels) thus curvatures and slopes, often of great interest in science, are given with good confidence. The proposed Virtual Image Correlation (VIC) method uses a virtual beam, an image which consists in a lateral expansion of the curve with a bell-shaped gray level. This figure is deformed until it fits the best the physical image with a method issued from the Digital Image Correlation method in use in solid mechanics. The precision of the identification is studied in a benchmark and successfully compared to two state of the art methods. Three practical examples are given: a bar bending under its own weight, a thin fiber transported by a flow within a fracture and a thermal front. The first allows a comparison with theoretical solution, the second shows the ability of the method to deal with complex shapes and crossings and the third deals with ill-defined image.

Virtual image correlation of magnetic resonance images for 3D geometric modelling of pelvic organs

“…Alternatively, a registration method called Virtual Image Correlation (VIC) has been proposed to detect objects in 2D images captured during mechanical experiments, using a deformable parametric geometry. [22,23] The virtual image is generated as an intensity profile linked with the geometry and is compared with the real image intensity. The correlated geometry is smooth and circumvents the deficiency related to image resolution or noises.…”

Deformation of a flexible fiber in a viscous flow past an obstacle

“…The process is repeated for several fiber lengths: 5 ≤ L f ≤ 80 mm. The values of δx, i.e., the vertical deflection of the tip of the fiber with respect to its position or a zero external load, and of L f are determined with an uncertainty of ±20 µm by means of an image analysis technique developed by Semin et al 24 In the present TABLE I. Physical and mechanical characteristics of the fibers (D, ρ, and E I are, respectively, the diameter (averaged over the length of the fibers and the orientation in a section), the density, and the bending stiffness of the fiber).…”

Multiorgan motion tracking in dynamic magnetic resonance imaging for evaluation of pelvic system mobility and shear strain

“…The levelset function defines the value of grey levels in the virtual image, which is similar to the ones in the real image. The use of virtual image was introduced in Semin et al and discussed in the literature . Now let τ be the neighbourhood width.…”