 This paper proposes a new approach to automatically reconstruct water regions from oblique photographs taken with unmanned aerial vehicles, UAV. First, it uses an algorithm to identify the edges of water regions in the photos. Then, it constructs a triangulated irregular network, TIN, of these identified edges. Next, it selects and processes textures from the photos to match them to the TIN. Finally, the textured TIN is mapped back onto the photo, resulting in a more accurate representation of the water region. This technique has been tested on real-world data and has proven successful at quickly and accurately reconstructing water regions in 3D models of oblique photography. This article was authored by Ingo Chio, Yakin Zhao, Jiuhua Luo, and others.