Solar supergranulation remains a mystery in spite of decades of intensive studies. Most of the papers about supergranulation deal with its surface properties. Local helioseismology provides an opportunity to look below the surface and see the vertical structure of this convective structure. We present a concept of a (3+1)-D segmentation algorithm capable of recognising individual supergranules in a sequence of helioseismic 3-D flow maps. As an example, we applied this method to the state-of-the-art data and derived descriptive statistical properties of segmented supergranules -- typical size of 20--30 Mm, characteristic lifetime of 18.7 hours, and estimated depth of 15--20 Mm. We present preliminary results obtained on the topic of the three-dimensional structure and evolution of supergranulation. The method has a great potential in analysing the better data expected from the helioseismic inversions, which are being developed.