The split-Lohmann lens technique has been proposed for computer-generated holograms, demonstrating exceptional computational efficiency. It requires a single 2D image, along with its corresponding depth map, to generate holograms with continuous depth at low computational cost. However, its reliance on a single-view depth map presents limitations, rendering it incapable of handling parallax and occlusion effects for multi-view observation. Polygon-based methods possess the capability to resolve this issue. This manuscript proposes an algorithm that applies the split-Lohmann lens technique to polygon-based holograms, demonstrating its feasibility for three-dimensional reconstruction and its superior computational efficiency compared to traditional polygon-based methods and the point-polygon hybrid method.