A method to characterize structure and symmetry in low-resolution images of colloidal thin films

A method is presented for characterizing particle centres, particle size and crystal symmetries with sub-pixel resolution from 8-bit digital images of colloidal thin films taken with a scanning electron microscope (SEM). Digital images are converted to xyz data points by converting colour contrast to a numerical intensity. The data are then passed through a modified form of a Savitzky–Golay filter which allows particle centres to be determined. A subsequent routine is presented that, by analysing the weighted standard deviation and average intensity of the pixels along shifting rings, improves the accuracy of the detected particle centres and provides the radius of each particle. Obtaining the particle centres allows the symmetry of each particle (with respect to its neighbours) along with the mean crystal orientation to be obtained, all in one cohesive package. A key advantage of the method presented here is that it is very robust and works with both low- and high-resolution images—enabling, for example, routine quantitative analysis of SEM images. Because of the low level of user input, the method can be used to process a batch of images in order to characterize the evolution of samples. (paper)