Roughness being a crucial feature of the surface texture is estimated through numerous techniques. The laser speckle imaging method has emerged as an efficient non-contact tool in the regime of surface roughness measurement techniques. This work presents singular value decomposition-based roughness measurement using objective speckle patterns of the machined surfaces. The surface roughness is quantified as a function of a proposed metric which is the exponential decay rate of the singular values associated with the speckle pattern. The effect of surface correlation length on the proposed metric is investigated and compared with speckle contrast and bright to dark pixel ratio of the binarized speckle pattern. The experimental results demonstrate the broad range surface roughness measuring capability of the proposed method using a single laser source.

粗糙度是表面纹理的一个重要特征，可以通过多种技术进行估计。激光散斑成像方法已成为表面粗糙度测量技术领域中一种有效的非接触式工具。这项工作使用加工表面的客观散斑图案提出了基于奇异值分解的粗糙度测量。表面粗糙度被量化为所提出的度量的函数，该度量是与散斑图案相关联的奇异值的指数衰减率。研究了表面相关长度对所提出的度量的影响，并与散斑对比度和二值化散斑图案的明暗像素比进行了比较。