Analysis of SEM digital images to quantify crack network pattern area in chromium electrodeposits

Abstract Scanning electron microscopy (SEM) is one of the primary methods used to inspect a wide range of materials, including metallic coatings. The use of software for image collection, analysis and, to some degree, interpretation is commonplace. However, no software seems to have been developed to quantify the cracks of chromium coatings. In the present paper SEM images are binarized using a proper threshold calculated from the histogram of images, once they have been filtered from nodules and background. The method uses the second derivative of the histogram obtained with the Laplacian of Gaussian (LoG), together with Prewitt vertical edge detector in some cases; then spatial cracked area can be quantified with much more accuracy than the subjective visual assessment that is usually done in the morphological analysis of the electrodeposits. The procedure has been automated so that the operator can straightforwardly obtain the result of the crack density from the original SEM image after some minor decisions orientated by the software, which is free on request from the authors. Manual computation has also proven to provide adequate results, but it takes longer and has limitations when highly cracked images are involved. The method could be probably extended to any kind of electrodeposits.