I converted the 217 photos I took around the statue into a point cloud and subsequently a textured high-poly mesh using Reality Scan. I then created a low-poly version of this high-poly mesh using Blender. I performed texture projection from the high-poly to the low-poly within Reality Scan. During this process, I corrected and cleaned up scanning errors caused by both the glare on the statue and the fountain underneath. Afterwards, I used AI to merge photos of the square from different angles to create an HDRI background image.

The limitations of using a phone camera and outdoor shooting are evident in the sharp-edged areas and the narrow sections of the statue. Sharp edges have glare while the scanning quality is lower in areas that can only be seen from limited angles.

Final model: 136k tris.