【opencv】图像畸变校正

接上篇文章：【鱼眼＋普通相机】相机标定

附代码：

方法一：

使用cv2.undistort

"""
    Create May 11, 2024
    @author Wang Jiajun
"""

import cv2
import numpy as np


def correct(img,camera_file='E:/calib_yaml/ship_calib.yaml'):
    """
    矫正图像，去除畸变
        img: 输入图像
        camera_file: 相机参数文件
        return: 矫正后的图像
    """
    fs = cv2.FileStorage(camera_file, cv2.FILE_STORAGE_READ)
    camera_matrix = fs.getNode("camera_matrix").mat()
    dist_coeffs = fs.getNode("dist_coeffs").mat()  # 畸变参数--顺序很重要！！！：k1,k2,p1,p2,k3
    resolution = fs.getNode("resolution").mat()  # 图像分辨率
    imSize = (resolution[0][0], resolution[1][0])
    fs.release()

    # 矫正图像
    img_distort = cv2.undistort(img, np.array(camera_matrix), np.array(dist_coeffs))
    return img_distort 


if __name__ == '__main__':
    camera_file = 'E:/calib_yaml/ship_calib.yaml'
    video = cv2.VideoCapture(0)
    while True:
        ret, img = video.read()
        # 3. 矫正图像
        img = cv2.resize(img, (640, 480))
        cv2.putText(img, "Press q to exit", (20, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 200, 0), 2)
        img_distort = cv2.correct(img, camera_file )
        # 4. 显示原图和矫正后的图像
        cv2.imshow('img_original', img)
        cv2.imshow('img_distort', img_distort)
        key = cv2.waitKey(1)
        # 5. 按q退出
        if key & 0xFF == ord('q'):
            break
    # 6. 关闭窗口
    video.release()
    cv2.destroyAllWindows()

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
40
41
42
43
44
45
46
47
48

方法二：

使用cv2.initUndistortRectifyMap建立映射关系，就不用每次读取内参了，速度更快。

"""
    Create May 11, 2024
    @author Wang Jiajun
"""

import cv2
import numpy as np


def correct(camera_file='E:/calib_yaml/ship_calib.yaml'):
    """
    矫正图像，去除畸变
        img: 输入图像
        camera_file: 相机参数文件
        return: 矫正后的图像
    """
    fs = cv2.FileStorage(camera_file, cv2.FILE_STORAGE_READ)
    camera_matrix = fs.getNode("camera_matrix").mat()
    dist_coeffs = fs.getNode("dist_coeffs").mat()  # 畸变参数--顺序很重要！！！：k1,k2,p1,p2,k3
    resolution = fs.getNode("resolution").mat()  # 图像分辨率
    imSize = (resolution[0][0], resolution[1][0])
    fs.release()

    # 矫正图像
    new_camera_matrix, roi = cv2.getOptimalNewCameraMatrix(np.array(camera_matrix), np.array(dist_coeffs),
                                                           imSize, 0,
                                                           imSize)
    map1, map2 = cv2.initUndistortRectifyMap(np.array(camera_matrix), np.array(dist_coeffs), None, new_camera_matrix,
                                             imSize, cv2.CV_32FC1)
    # img_distort = cv2.undistort(img, np.array(camera_matrix), np.array(dist_coeffs))
    return map1, map2


if __name__ == '__main__':
    camera_file = 'E:/calib_yaml/ship_calib.yaml'
    map1, map2 = correct(camera_file)
    video = cv2.VideoCapture(0)
    while True:
        ret, img = video.read()
        # 3. 矫正图像
        img = cv2.resize(img, (640, 480))
        cv2.putText(img, "Press q to exit", (20, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 200, 0), 2)
        img_distort = cv2.remap(img, map1, map2, cv2.INTER_LINEAR)
        # 4. 显示原图和矫正后的图像
        cv2.imshow('img_original', img)
        cv2.imshow('img_distort', img_distort)
        key = cv2.waitKey(1)
        # 5. 按q退出
        if key & 0xFF == ord('q'):
            break
    # 6. 关闭窗口
    video.release()
    cv2.destroyAllWindows()

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
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54