Evaluation methods for gradient measurement techniques

Many optical metrology methods deliver 2D fields of gradients, such as shearography, Shack-Hartmann sensors and the fringe reflection technique that produce gradients for deformation, wave-front shape and object shape, respectively. The evaluation for gradient data usually includes data processing, feature extraction and data visualization. The matters of this talk are optimized and robust processing methods to handle and prepare the measured gradients. Special attention was directed to the fact that optical measurements typically produce data far from ideal behavior and that parts of the measured area are usually absent or invalid. A robust evaluation must be capable to deliver reliable results with non perfect data and the evaluation speed should be sufficient high for industrial applications. Possible data analysis methods for gradients are differentiation and further integration as well as vector processing when orthogonal gradients are measured. Evaluation techniques were investigated and optimized (e.g. for effective bump and dent analysis). Key point of the talk will be the optimized data integration that delivers the potential of measured gradients. I.e. for the above mentioned examples: the deformation, wave-front and object shape are delivered by successful data integration. Local and global existing integration methods have been compared and the optimum techniques were combined and improved for an accelerated and robust integration technique that is able to deal with complicated data validity masks and noisy data with remaining vector rotation which normally defeats a successful integration. The evaluation techniques are compared, optimized and results are shown for data from shearography and the fringe reflection technique (, which is demonstrated in talk “High Resolution 3D Shape Measurement on Specular Surfaces by Fringe Reflection”).