The process of ice hummock formation under pressure was simulated using a two-dimensional stand. The purpose of the work was to observe the movement of polypropylene plates, which simulate ice blocks, relative to each other during the simulation of an ice hummock formation. The main idea of the stand is to limit the movement of the ice simulating blocks only in the vertical plane. This plane creates the illusion of a cross-section of a hummock. The design of the stand and the modeling methodology are discussed. The process of physical modeling was recorded on video. Using computer vision technology, each frame was analyzed based on operations with vector polygons. As a result of image processing, 20 morphometric parameters of the hummock structure were recorded, including: sail width, keel width, coordinates of the upper point of the sail, coordinates of the lower point of the keel, sail, keel, area of blocks, distribution of porosity horizontally and vertically, position of barycenters, etc. Keel draft of the model intensifies as the total area of the blocks involved into the experiment increases. The increase in the model is proportional to the square root of the total area of the blocks with a coefficient of 0.8. The obtained model cross-section profiles were compared with the real cross-sections of the ice hummocks. The results are quite satisfactory, indicating that the modeling adequately reflects the formation of real ice hummocks. In a number of experiments, blocks of two colors were used consistently, and the total areas of blocks of both colors were the same. When compressed, the impending thin ice goes up, rafting and subsequently collapsing. The introduction of blocks into the forming hummock occurs both from above in the form of a layering, and inside the keel in the form of “jets”. Further, the blocks under the influence of gravity move downwards, forming a kind of whirlwind, thereby determining the predominant scenario of hummocking. According to the second scenario, which can be called the “adjoining” scenario, the accumulation of new blocks occurs at the edge of the formed keel, and no layering or subsequent swirls take place. The second scenario of hummocking occurs much less frequently, in approximately 20% of the experiments. At the moment, the reasons for hummocking in either scenario remain unclear.