Image analysis is an efficient technique used in many areas of science and industry. However, in process analytical applications it tends to be an ancillary tool, used mainly for visual monitoring or measuring some geometrical properties. At the same time, there are many other important aspects of the process samples appearance, besides measurable distances, that may be connected to the information of interest. In the present paper, the methods of image analysis were applied to at-line monitoring of fluid bed pellet coating process. The quantitative description of images of pellet samples, taken from different process stages, has been obtained using two different approaches: wavelet decomposition and angle measure technique (AMT). Both methods revealed a strong correlation between image features and process parameters. However, the AMT results turned out to be more accurate and stable. It has been shown that pellet images, taken with a conventional digital camera, can be used for at-line monitoring of the process course, specifically, the growth of pellets due to the coating. An algorithm for precise counting of pellets has been developed. Combined with the sample weighing, it enables an accurate determination of the mean added pellets' weight. The method can be used for the determination of the mean layer thickness, either by itself for at-line analysis or as a reference technique, when modeling the process from in-line spectroscopy data.