3D Scanning

I tried out a few applications on my iphone that use photogrammetry technology.

FOCOS captures a 2.5D information of the picture by making a depth map with iphone's dual camera. The depth map makes it able to pick focal point after the photo has been taken, add or adjust the light condition in 3D space, etc.

I found a really nice video tutorial that covers every step of installing ffmpeg on Windows! You can check on it here.

SCANDY seems to be designed solely for facial scanning. It only works with a selfie camera, and doesn't even allow switching the camera. It takes pictures and builds mesh out of them in real time. It actually does it pretty fast; you can get a full 3D model of your face in less than 2 minutes. You can export the scanned file into OBJ, STL, etc.

Building 3D Mesh From Video

Then, I wanted to scan something big, and started to look for other photogrammetry softwares. VisualSFM is an open-source GUI application for 3D reconstruction using structure from motion. I found nice tutorial from Youtube VisualSFM locates the position of the camera at each point by matching images at each point, and maps point clouds from the images. I took a video around a huge Taihu stone, exhibited in the front of Boston Museum of Fine arts. Then I extracted a hundred thumbnails from the video.

Due to the bug while rebuilding a mesh out of the point cloud, I tried to use other programs. First I tried Meshroom, but found out it somehow requires NVIDIA GPU to build mesh out of point cloud. Then I moved on to Metashape, which finally had some good result in producing mesh out of the thumbnails. Below is the final result of the scanning.

Then I printed using Sindoh with minimum wall thickness of 0.15mm.