使用Python进行相机校准 – OpenCV

先决条件： OpenCV

相机是机器人、太空探索等多个领域不可或缺的一部分，相机发挥着重要作用。它有助于捕捉每一刻，并有助于许多分析。为了将相机用作视觉传感器，我们应该知道相机的参数。相机校准只不过是估计相机的参数，需要有关相机的参数来确定现实世界中的 3D 点与其在该校准相机捕获的图像中对应的 2D 投影（像素）之间的准确关系。

我们需要考虑镜头的焦距、光学中心和径向畸变系数等内部参数，以及相机相对于某些现实世界坐标系的旋转和平移等外部参数。

需要的库：

OpenCV Python中的库是一个计算机视觉库，主要用于图像处理、视频处理和分析、人脸识别和检测等。
麻木是一个通用的数组处理包。它提供了一个高性能的多维数组对象和用于处理这些数组的工具。

相机校准可以分步完成：

步骤 1：首先使用已知大小的棋盘图案定义 3D 点的真实世界坐标。
步骤2：捕获棋盘图像的不同视点。
第3步： findChessboardCorners()是OpenCV中的一个方法，用于寻找不同图像中每个3D点的像素坐标(u,v)
步骤 4：然后使用calibrateCamera()方法查找相机参数。

它将以我们计算的（threedpoints、twodpoints、grayColor.shape[::-1]、None、None）作为参数，并返回包含相机矩阵、失真系数、旋转向量和平移向量等元素的列表。

相机矩阵有助于将 3D 对象点转换为 2D 图像点，并且失真系数返回相机在世界中的位置，以及旋转和平移向量的值

以下是上述方法的完整程序：

Python3

# Import required modules
import cv2
import numpy as np
import os
import glob
 
 
# Define the dimensions of checkerboard
CHECKERBOARD = (6, 9)
 
 
# stop the iteration when specified
# accuracy, epsilon, is reached or
# specified number of iterations are completed.
criteria = (cv2.TERM_CRITERIA_EPS +
            cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)
 
 
# Vector for 3D points
threedpoints = []
 
# Vector for 2D points
twodpoints = []
 
 
#  3D points real world coordinates
objectp3d = np.zeros((1, CHECKERBOARD[0]
                      * CHECKERBOARD[1],
                      3), np.float32)
objectp3d[0, :, :2] = np.mgrid[0:CHECKERBOARD[0],
                               0:CHECKERBOARD[1]].T.reshape(-1, 2)
prev_img_shape = None
 
 
# Extracting path of individual image stored
# in a given directory. Since no path is
# specified, it will take current directory
# jpg files alone
images = glob.glob('*.jpg')
 
for filename in images:
    image = cv2.imread(filename)
    grayColor = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
 
    # Find the chess board corners
    # If desired number of corners are
    # found in the image then ret = true
    ret, corners = cv2.findChessboardCorners(
                    grayColor, CHECKERBOARD,
                    cv2.CALIB_CB_ADAPTIVE_THRESH
                    + cv2.CALIB_CB_FAST_CHECK +
                    cv2.CALIB_CB_NORMALIZE_IMAGE)
 
    # If desired number of corners can be detected then,
    # refine the pixel coordinates and display
    # them on the images of checker board
    if ret == True:
        threedpoints.append(objectp3d)
 
        # Refining pixel coordinates
        # for given 2d points.
        corners2 = cv2.cornerSubPix(
            grayColor, corners, (11, 11), (-1, -1), criteria)
 
        twodpoints.append(corners2)
 
        # Draw and display the corners
        image = cv2.drawChessboardCorners(image,
                                          CHECKERBOARD,
                                          corners2, ret)
 
    cv2.imshow('img', image)
    cv2.waitKey(0)
 
cv2.destroyAllWindows()
 
h, w = image.shape[:2]
 
 
# Perform camera calibration by
# passing the value of above found out 3D points (threedpoints)
# and its corresponding pixel coordinates of the
# detected corners (twodpoints)
ret, matrix, distortion, r_vecs, t_vecs = cv2.calibrateCamera(
    threedpoints, twodpoints, grayColor.shape[::-1], None, None)
 
 
# Displaying required output
print(" Camera matrix:")
print(matrix)
 
print("\n Distortion coefficient:")
print(distortion)
 
print("\n Rotation Vectors:")
print(r_vecs)
 
print("\n Translation Vectors:")
print(t_vecs)

输入：

输出：

Camera matrix: 
[[ 36.26378216   0.         125.68539168]
 [  0.          36.76607372 142.49821147]
 [  0.           0.           1.        ]]

 Distortion coefficient: 
[[-1.25491812e-03  9.89269357e-05 -2.89077718e-03  4.52760939e-04
  -3.29964245e-06]]

 Rotation Vectors: 
[array([[-0.05767492],
       [ 0.03549497],
       [ 1.50906953]]), array([[-0.09301982],
       [-0.01034321],
       [ 3.07733805]]), array([[-0.02175332],
       [ 0.05611105],
       [-0.07308161]])]
 
 Translation Vectors: 
[array([[ 4.63047351],
       [-3.74281386],
       [ 1.64238108]]), array([[2.31648737],
       [3.98801521],
       [1.64584622]]), array([[-3.17548808],
       [-3.46022466],
       [ 1.68200157]])]