File size: 1,151 Bytes
c608946
 
 
 
 
c74a070
c608946
 
 
 
c74a070
c608946
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c74a070
c608946
c74a070
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
from PIL import Image
import torch
import cv2
from roma import roma_outdoor

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")


if __name__ == "__main__":
    from argparse import ArgumentParser

    parser = ArgumentParser()
    parser.add_argument("--im_A_path", default="assets/sacre_coeur_A.jpg", type=str)
    parser.add_argument("--im_B_path", default="assets/sacre_coeur_B.jpg", type=str)

    args, _ = parser.parse_known_args()
    im1_path = args.im_A_path
    im2_path = args.im_B_path

    # Create model
    roma_model = roma_outdoor(device=device)

    W_A, H_A = Image.open(im1_path).size
    W_B, H_B = Image.open(im2_path).size

    # Match
    warp, certainty = roma_model.match(im1_path, im2_path, device=device)
    # Sample matches for estimation
    matches, certainty = roma_model.sample(warp, certainty)
    kpts1, kpts2 = roma_model.to_pixel_coordinates(matches, H_A, W_A, H_B, W_B)
    F, mask = cv2.findFundamentalMat(
        kpts1.cpu().numpy(),
        kpts2.cpu().numpy(),
        ransacReprojThreshold=0.2,
        method=cv2.USAC_MAGSAC,
        confidence=0.999999,
        maxIters=10000,
    )