116 lines
3.8 KiB
Python
116 lines
3.8 KiB
Python
import cv2
|
|
import numpy as np
|
|
import plotly.express as px
|
|
import plotly.io as pio
|
|
|
|
pio.renderers.default = "browser"
|
|
|
|
|
|
def read_camera_parameters(file_path):
|
|
cameras = []
|
|
with open(file_path, "r") as f:
|
|
n_cameras = int(f.readline())
|
|
for _ in range(n_cameras):
|
|
line = f.readline().split()
|
|
camera = {
|
|
"image": line[0],
|
|
"K": np.array(
|
|
[
|
|
[float(line[1]), 0, float(line[3])],
|
|
[0, float(line[5]), float(line[6])],
|
|
[0, 0, 1],
|
|
]
|
|
),
|
|
"R": np.array(
|
|
[
|
|
[float(line[10]), float(line[11]), float(line[12])],
|
|
[float(line[13]), float(line[14]), float(line[15])],
|
|
[float(line[16]), float(line[17]), float(line[18])],
|
|
]
|
|
),
|
|
"t": np.array(
|
|
[[float(line[19])], [float(line[20])], [float(line[21])]]
|
|
),
|
|
}
|
|
cameras.append(camera)
|
|
return cameras
|
|
|
|
|
|
def reconstruct_3d_points(img1, img2, K, R1, t1, R2, t2):
|
|
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
|
|
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
|
|
sift = cv2.SIFT_create()
|
|
kp1, des1 = sift.detectAndCompute(gray1, None)
|
|
kp2, des2 = sift.detectAndCompute(gray2, None)
|
|
bf = cv2.BFMatcher(cv2.NORM_L2, crossCheck=True)
|
|
matches = bf.match(des1, des2)
|
|
matches = sorted(matches, key=lambda x: x.distance)
|
|
N_BEST_MATCHES = 100
|
|
matches = matches[:N_BEST_MATCHES]
|
|
pts1 = np.float32([kp1[m.queryIdx].pt for m in matches]).reshape(-1, 1, 2)
|
|
pts2 = np.float32([kp2[m.trainIdx].pt for m in matches]).reshape(-1, 1, 2)
|
|
P1 = K @ np.hstack((R1, t1))
|
|
P2 = K @ np.hstack((R2, t2))
|
|
E, mask = cv2.findEssentialMat(
|
|
pts1, pts2, K, method=cv2.RANSAC, prob=0.999, threshold=1.0
|
|
)
|
|
matches = [m for i, m in enumerate(matches) if mask[i][0]]
|
|
pts1 = pts1[mask.ravel() == 1]
|
|
pts2 = pts2[mask.ravel() == 1]
|
|
points_4D = cv2.triangulatePoints(P1, P2, pts1, pts2)
|
|
points_3D = points_4D / points_4D[3]
|
|
points_3D = points_3D[:3].T
|
|
img_matches = cv2.drawMatches(
|
|
img1,
|
|
kp1,
|
|
img2,
|
|
kp2,
|
|
matches,
|
|
None,
|
|
flags=cv2.DrawMatchesFlags_NOT_DRAW_SINGLE_POINTS,
|
|
)
|
|
cv2.imshow("Matches", img_matches)
|
|
cv2.waitKey(1)
|
|
return points_3D
|
|
|
|
|
|
def main():
|
|
cameras = read_camera_parameters(r"10\templeRing\templeR_par.txt")
|
|
all_points_3D = []
|
|
for i in range(len(cameras) - 1):
|
|
print(f"Processing image pair {i+1}/{len(cameras)-1}")
|
|
img1 = cv2.imread(f"10/templeRing/{cameras[i]['image']}")
|
|
img2 = cv2.imread(f"10/templeRing/{cameras[i+1]['image']}")
|
|
if img1 is None or img2 is None:
|
|
print(f"Failed to load images {i} and {i+1}")
|
|
continue
|
|
points_3D = reconstruct_3d_points(
|
|
img1,
|
|
img2,
|
|
cameras[i]["K"],
|
|
cameras[i]["R"],
|
|
cameras[i]["t"],
|
|
cameras[i + 1]["R"],
|
|
cameras[i + 1]["t"],
|
|
)
|
|
distances = np.sqrt(np.sum(points_3D**2, axis=1))
|
|
mask = distances < np.percentile(distances, 95)
|
|
points_3D = points_3D[mask]
|
|
all_points_3D.append(points_3D)
|
|
cv2.destroyAllWindows()
|
|
all_points_3D = np.vstack(all_points_3D)
|
|
fig = px.scatter_3d(
|
|
x=all_points_3D[:, 0],
|
|
y=all_points_3D[:, 1],
|
|
z=all_points_3D[:, 2],
|
|
title="3D Reconstruction from Multiple Views",
|
|
)
|
|
fig.update_traces(marker=dict(size=1))
|
|
fig.update_layout(scene=dict(aspectmode="data"))
|
|
fig.show()
|
|
fig.write_html("10/templeR_multi.html")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|