Barley primarily serves as a major animal feed crop; smaller amounts of barley are used in health foods and in the malting process. Detailed geometric parameters of kernels are very important for the design of food engineering processes, such as the air transport, drying, milling, and malting. Image analysis was used to determine the size parameters of one hundred kernels of selected varieties of <I>Hordeum vulgare</I> L. The data for every kernel captured were stored for further use, together with the mean, standard deviation (SD), coefficient of variation (<I>CV</I>), and images themselves. The measured data were then used to compute the volume and surface area of each of the five kernel models (Models 0–4), the results being subsequently verified by pycnometric measurement. Model 0 represents the general ellipsoid, models 1–3 various combinations of two parts of a general ellipsoid with one or two cone frustums. The best fitted model 4 was a combination of two cone frustums. Based on the results of image analysis measurements and on the presented model 4, a simplified method for the specific surface estimation of barley grains from the weight of 1000 kernels is recommended.