A personal image-stitching algorithm implementation for the Image Processing laboratory. Uses the Waymo Public Dataset.
To obtain the panoramic images you should have a folder named dataset/camera_image where you should load the .parquet files. Also very important for the semantic panoramas you should have a folder dataset/semantic/camera_segmentation and also the result forlder semantic_results.
To create both the semantic and normal panoramas you should give the flaot('inf') argument of the main() function inside the main.py. If youn want just the semantic give the argument 1. To run create a video from the panorama images you should change the target directories inside the video.py file then run the command python video.py.