The performance of diode laser based moire deflectometry and video projection moire techniques will be compared for use in optical remote sensing of fatigue cracks in highway bridges. The sensing method will be evaluated using painted steel samples undergoing cyclic dynamic loading to induce fatigue cracking under controlled conditions. The loading cycle will be chosen to mimic that of normal traffic flow. The techniques are designed to sense characteristic surface deformations in the vicinity of the crack tip. Deflectometry is sensitive to local surface slopes, while projection techniques map surface height contours. A variety of signal processing techniques will be used to enhance defect visibility and quantification including Fourier filtering of static images, and differencing and error mapping techniques using sequential images obtained at varying load. By using naturally occurring "markers" on the surface, we hope to use difference imaging techniques to gain information about in-plane deformations. Error mapping techniques can be used to image only the changes in the moire pattern during loading. Characteristic deformation pattern images will form the basis for autonomous, trainable defect identification systems using neural networks to be developed during the Phase II effort.Anticipated Results/Potential Commercial Applications of Results:At the close of the anticipated Phase II program we will have demonstrated a prototype mobile remote sensing system for detection of fatigue cracks. this system will provide the basis for development of trainable, autonomous defect sensors for infrastructure integrity evaluations and real-time manufacturing inspection applications. Manufacturing applications include flaw detection on ground surfaces, crack detection in stamped parts, surface finish inspection, identification of defects in glass parts, reading of embossed part identifications, and IC chip loading guidance and defect identification in printed circut boards
