hynpu/surround-view-system-introduction — reverse-engineered prompt

Build me a simple Python surround view camera system like the one in this repo. I want it to help a beginner understand the full workflow from camera calibration, to generating projection maps, to creating blend weight masks, to stitching the camera views together, and then running a live demo.

Please make it practical rather than overly fancy. It should have clear scripts for each step, a folder for config files, sample images or placeholders where needed, and readable documentation explaining how to use it. The final result should be able to take multiple camera feeds, project them into a bird's eye view, blend the overlaps smoothly, and show the combined surround view in real time if cameras are available.

Keep the code easy to follow, with comments where the math or image processing might be confusing. If you need exact OpenCV details or calibration steps, look up current docs online.