### 使用OpenCV进行鱼眼图像畸变校正 #### 准备工作 为了实现鱼眼图像的畸变校正，需要准备一系列带有棋盘格图案的照片用于相机标定。通过这些照片可以计算出相机内部参数矩阵以及畸变系数。 #### 寻找角点 调用`cv2.findChessboardCorners()`函数来检测每张输入图片上是否存在标准尺寸的棋盘格模式，并获取其二维位置信息[^1]。 ```python import cv2 import numpy as np chessboard_size = (7, 6) # 棋盘格大小 objp = np.zeros((np.prod(chessboard_size), 3), dtype=np.float32) objp[:, :2] = np.mgrid[0:chessboard_size[0], 0:chessboard_size[1]].T.reshape(-1, 2) img_points = [] # 存储所有找到的图像空间内的角点坐标 obj_points = [] # 对应的实际世界中的物体坐标系下的坐标 for fname in images: img = cv2.imread(fname) gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) ret, corners = cv2.findChessboardCorners(gray, chessboard_size, None) if ret: obj_points.append(objp) refined_corners = cv2.cornerSubPix( gray, corners, (11, 11), (-1, -1), criteria=(cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)) img_points.append(refined_corners) ``` #### 鱼眼镜头标定 当收集足够的有效样本后，就可以使用`cv2.fisheye.calibrate()`完成对鱼眼镜头的具体参数估计过程，得到内外参数据。 ```python N_OK = len(obj_points) K = np.zeros((3, 3)) # 内参矩阵 D = np.zeros((4, 1)) # 畸变向量 rvecs = [np.zeros((1, 1, 3), dtype=np.float64) for i in range(N_OK)] # 旋转矢量列表 tvecs = [np.zeros((1, 1, 3), dtype=np.float64) for i in range(N_OK)] # 平移矢量列表 rms, _, _, _, _ = \ cv2.fisheye.calibrate( obj_points, img_points, gray.shape[::-1], K, D, rvecs, tvecs, calibration_flags=cv2.fisheye.CALIB_RECOMPUTE_EXTRINSIC + cv2.fisheye.CALIB_CHECK_COND + cv2.fisheye.CALIB_FIX_SKEW, criteria=(cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 1e-6)) print("Found " + str(rms) + " root mean square error") print("\nIntrinsic matrix:\n" + str(K)) print("\nDistortion coefficients:\n" + str(D.ravel())) ``` #### 图像去畸变 有了上述获得的内参和畸变参数之后，便可以通过`cv2.fisheye.undistortImage()`方法去除原始拍摄所得图像中存在的桶形失真现象。 ```python def undistort_fisheye_image(image_path, output_path, camera_matrix, dist_coeffs): image = cv2.imread(image_path) h, w = image.shape[:2] map1, map2 = cv2.fisheye.initUndistortRectifyMap( camera_matrix, dist_coeffs, R=np.eye(3), P=cv2.getOptimalNewCameraMatrix(camera_matrix, dist_coeffs, (w, h), 1, (w, h))[0], size=(w, h), m1type=cv2.CV_32FC1) undistorted_img = cv2.remap(image, map1, map2, interpolation=cv2.INTER_LINEAR, borderMode=cv2.BORDER_CONSTANT) cv2.imwrite(output_path, undistorted_img) image_to_correct = 'path/to/fisheye/image.jpg' corrected_image_output = 'output/corrected_image.png' undistort_fisheye_image(image_to_correct, corrected_image_output, K, D) ```