Inverting a real-valued index grid

Question

OpenCV\'s remap() uses a real-valued index grid to sample a grid of values from an image using bilinear interpolation, and returns the grid of samples as a new image.

<

Answer 1

You can invert map at known points and interpolate it into new grid. It will work fine, while distortion is not very huge.

Here is very simple implementation in Python using scipy.interpolate.griddata:

map_x, map_y = cv2.initUndistortRectifyMap(K, D, None, new_K, image_size, cv2.CV_32FC1)

points =  np.stack([map_x.flatten(), map_y.flatten()], axis=1)
grid = np.mgrid[:map_x.shape[0], :map_y.shape[1]]
values = grid.reshape(2, -1).T[..., ::-1] 

from scipy.interpolate import griddata
grid_y, grid_x = grid
map_back = griddata(points, values, (grid_x, grid_y), method='cubic').astype(map_undistort.dtype)

If you use CV_32FC2 for maps, you can simplify points construction:

map_undistort, _ = cv2.initUndistortRectifyMap(K, D, None, new_K, image_size, cv2.CV_32FC2)
points = map_undistort.reshape(-1, 2)