OpenCV 4 with Python Blueprints - Second Edition

Applying image rectification

Maybe the easiest way to make sure that we have recovered the correct camera matrices is to rectify the images. If they are rectified correctly, then a point in the first image and a point in the second image that corresponds to the same 3D world point will lie on the same vertical coordinate.

In a more concrete example, such as in our case, since we know that the cameras are upright, we can verify that horizontal lines in the rectified image correspond to horizontal lines in the 3D scene. Thus, we follow these steps to rectify our image:

1. First, we perform all the steps described in the previous subsections to obtain the [R | t] matrix of the second camera:

def plot_rectified_images(self, feat_mode="SIFT"): ...