Imaging Surface Roughness using Correlated Speckle Grain Pairs

This paper reports on a laser speckle technique based on that which was first developed by Russo [1] to measure surface roughness. The technique has been extended to image surface roughness using laser scanning. A one dimensional laser scan is used with the second dimension required for imaging provided by movement of the surface. The technique uses two orthogonally polarised laser beams at a slight relative angle to each other to produce two independent speckle patterns that are imaged and auto-correlated to deduce the roughness of the surface being illuminated. The offset angle between the beams is a plane orthogonal to the scan direction and is relatively insensitive to scan position. Results to date are promising but show an unwanted variation of the auto-correlation value with the average brightness of the surface. This paper discusses the method, results and outlines further work required.