An integral imaging system based on hierarchical clustering two-dimensional viewpoint synthesis is proposed. In this integral imaging system, a sparse viewpoint image is collected using a motorized translation stage combined with a single camera viewpoint collection method. In order to improve the efficiency of viewpoint collection, a two-dimensional virtual viewpoint synthesis method based on hierarchical clustering is used to realize virtual viewpoint synthesis between images, and the synthesized multi-viewpoint image is used to obtain an elemental image array through pixel mapping. By using pixel mapping to obtain elemental image array, the computational reconstruction process is achieved. Experimental results show that the system can not only meet the high-efficiency three-dimensional information collection, but also improve the resolution quality and increase the application flexibility while ensuring quality of the synthetic viewpoint.