import numpy as np import cv2 import glob # 1. 准备标定板参数 chessboard_size = (7, 10) # 内角点数量 square_size = 13.0 # 毫米 # 2. 生成世界坐标系中的3D点 objp = np.zeros((chessboard_size[0]*chessboard_size[1], 3), np.float32) objp[:, :2] = np.mgrid[0:chessboard_size[0], 0:chessboard_size[1]].T.reshape(-1, 2) * square_size # 3. 存储对象点和图像点 objpoints = [] # 3D点 imgpoints = [] # 2D点 # 4. 读取所有标定图像 images = glob.glob('D:/DeepLearning/datasets/clibration/rect/*.png') image_size = None for fname in images: img = cv2.imread(fname) gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) if image_size is None: image_size = gray.shape[::-1] # 查找棋盘格角点 ret, corners = cv2.findChessboardCorners(gray, chessboard_size, None) if ret: objpoints.append(objp) # 亚像素精确化 corners2 = cv2.cornerSubPix(gray, corners, (11,11), (-1,-1), (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)) imgpoints.append(corners2) # 可视化角点 cv2.drawChessboardCorners(img, chessboard_size, corners2, ret) cv2.imshow('Corners', img) cv2.imwrite("this.png",img) cv2.waitKey(500) cv2.destroyAllWindows() # 5. 相机标定 ret, mtx, dist, rvecs, tvecs = cv2.calibrateCamera(objpoints, imgpoints, image_size, None, None) print("相机内参矩阵:\n", mtx) print("\n畸变系数:", dist.ravel()) # 6. 评估标定误差 mean_error = 0 for i in range(len(objpoints)): imgpoints2, _ = cv2.projectPoints(objpoints[i], rvecs[i], tvecs[i], mtx, dist) error = cv2.norm(imgpoints[i], imgpoints2, cv2.NORM_L2)/len(imgpoints2) mean_error += error print("\n平均重投影误差: {} 像素".format(mean_error/len(objpoints))) # 7. 保存标定结果 np.savez('calibration_result.npz', mtx=mtx, dist=dist)