Measurement of surface roughness over a large area is a very challenging task due to the limitations with the existing techniques. Surface roughness measurement techniques including stylus and microscopy are limited by point-by-point data acquisition and a small field of view (FOV). In effect, any solution that would subdue these limitations would be characterized by its full-field nature, large FOV, and the ability to acquire and process data at high speeds. To meet these requirements, large area speckle imaging has been used to obtain areal surface roughness parameters through the processing of spectrally correlated speckle images. An automated optical system is developed for surface roughness evaluation of components with large and curved surface areas. In order to extract areal surface roughness parameters from the captured set of images, processing algorithms are developed. The methodology is first validated using a comparator plate containing areas having an average surface roughness (Ra) ranging between 0.2 µm and 0.6 µm. Further, statistical significance tests are conducted to determine the main factors affecting system performance. The measurement results are compared and validated using a 3D optical microscope. The results obtained from the blind tests performed on aerospace component surfaces as large as 450mm×210mm are also presented.