At a basic level, manufactured surfaces can be evaluated according to properties including their primary geometry or shape (topography), waviness, and varying degrees of structures and roughness (2D/3D). Each of these surface features contain both intentional and unintentional (controlled/uncontrolled) contributions. Depending on the context; the structure, waviness, and roughness contributions might be categorized as finish, texture, defects, marks, waves or one of many other designations.
The objective is to evaluate (qualify and quantify) the different properties of a surface in accordance with the industry's set targets. The used technology is based on the physical principle of deflectometry. An optical, contact-less solution to detect, characterize and classify the entire range of low and high frequency defects. Thus, providing means to assist or replace human subjectivity.
and derivative using specific mathematical algorithms
Slope gradient analysis provides a precise cartography of the measured surface
Deflectometry analysis can either be applied through a temporal (phase-shifting) mode or through a spatial mode, the latter method requiring only a single image. The applied principle varies in relation with the environment and needed spatial resolution.
Principles of measurement:
The applied technology is based on advanced reflective Deflectometry. The measurement setup requires a dynamic screen to display periodic fringes using structured light. As well as a high resolving camera set observing the reflected light on the measured surface (near the center of curvature of the tested sample).
Since light rays are reflected by the surface with an angle of incidence equal to the angle of reflection. The measured surface slope will be distorted where a local slope variation suddenly modifies the path of the reflected image. Deflectometry provides an accurate solution to measure the local slope map and, with the help of digital derivation, enables the quantification of the local curvature variations.
A possible surface defect is described as a rapid and sudden change in the surface’s local slope. The relevant information for defect visualization and quantification is found in the slopes’ gradients.
The "reflected fringe technique" is used to gauge objects with specular reflective surfaces, with the specular surface acting as mirror. A computer-generated fringe pattern is displayed (using a dynamic screen), the pattern being reflected on the inspected surface. The reflection feeds a virtual image of the fringe pattern to a dedicated camera, digitalizing the image as a result. The acquired fringe pattern reflection will be distorted according to changes in the slope of the measured surface, highlighting quality defects.
Phase Shifting Deflectometry:
Phase shifting deflectometry provides a way to measure the local slope map of a surface and, through digital derivation, enables the quantification of the local curvature variations responsible for the quality defects on the surface.
The displayed pattern is shifted in X & Y direction in synchronization with the camera acquisition of the reflected images.
The reflected images generated through phase shifting are processed to obtain a phase map which is then unwrapped to obtain a final slope map. Thus, providing a first level of measurement. Using different methods, the slope map can be further derived to obtain the curvature map, then processed to finally extract the relevant information of the wavelength corresponding to the sought defect.
In the last processing step, high performance image analysis is used to quantify and classify the extracted information into a comprehensive end-result report.
Sample in glass fragmentation testing
You may consult the OPTIFRAG product page for further details on Image acquisition, boundaries detection & glass fragment labeling.
2. Detected fragment boundaries highlighted in red
1. Raw acquisition of the reflected pattern image
3. Labeling map highlighting individual fragments
N.B.: The 2 horizontal defrosting wires in the raw image (1.) are not detected as boundaries in (2.) and thus don’t generate false fragments in (3.)