OpenCV 4 with Python Blueprints - Second Edition

Creating the canvas for the panorama

Now it's time to create the canvas. For that, we create a warper object based on our desired rotation schema. For simplicity, let's assume a planar model:

    warper = cv2.PyRotationWarper('plane', 1)

Then, we enumerate over all the connected images and get all the regions of interest in each of the images:

    stitch_sizes, stitch_corners = [], []    for i, img in enumerate(conn_images):        sz = img.shape[1], img.shape[0]        K = cameras[i].K().astype(np.float32)        roi = warper.warpRoi(sz, K, cameras[i].R)        stitch_corners.append(roi[0:2])        stitch_sizes.append(roi[2:4])

Finally, we estimate the final canvas_size based on all regions of interest:

 canvas_size = cv2.detail.resultRoi(corners=stitch_corners, sizes=stitch_sizes) ...