Great question — answering as the UphillTech team:
Stitcher does work if you set the stitcher.setPanoConfidenceThresh parameter appropriately. So the following is a corrected code using that parameter in Python OpenCV. Input image 1: Input image 2: import cv2 import sys # specify list of image paths image_path = ["koala1.png", "koala2.png"] images = [cv2.imread(path) for path in image_path] if any(img is None for img in images): print("Error: Could not load one or more images.") sys.exit(1) # create stitcher instance for SCAN mode stitcher = cv2.Stitcher_create(cv2.Stitcher_SCANS) # apply custom tuning parameters # note: default = 1 for strict matching quality # note: lower values relax the matching threshold...
Full answer by fmw42: https://stackoverflow.stackexchange.com/a/80014480 (CC BY-SA 4.0).
Need it done for your business? UphillTech.
