OpenCV下的无标定校正(stereoRectifyUncalibrated)

QuanHaHQuan

已于 2024-09-03 18:07:19 修改

阅读量1.1k

点赞数 6

CC 4.0 BY-SA版权

分类专栏： Blah-Blah code sharing 文章标签： opencv 人工智能计算机视觉 python

于 2024-09-03 18:06:44 首次发布

本文链接：https://blog.youkuaiyun.com/c_h_q_/article/details/141868270

OpenCV下的无标定校正(stereoRectifyUncalibrated)

文章目录

1. 杂话
2. 无标定校正
3. 矫正结果

1. 杂话

咱们在之前的帖子里面讲了一些比较常规的标定和校正OpenCV下的单目标定，双目标定与立体校正(calibrateCamera, stereoCalibrate and stereoRectify)，但是，如果说两个相机之间的转角比较大，那么可能并不适合用双目标定＋立体校正的步骤来做，参考这个贴子立体校正失效的情况，所以为了解决这个问题，今天咱们就来说说如何使用非标定的校正函数：

对了，我使用的数据和完整的代码在这里：Repo : Calibrate-and-Rectify

2. 无标定校正

2.1 先看代码

import cv2
import numpy as np
import glob
import matplotlib.pyplot as plt
from PIL import Image

c1_images_names = sorted(glob.glob('demo/left*.jpg'))
c2_images_names = sorted(glob.glob('demo/right*.jpg'))
 
c1_images = []
c2_images = []
for im1, im2 in zip(c1_images_names, c2_images_names):
    _im = cv2.imread(im1, 1)
    c1_images.append(_im)
 
    _im = cv2.imread(im2, 1)
    c2_images.append(_im)

def find_corners(image_names):
    images = []
    for imname in image_names:
        im = cv2.imread(imname, 1)
        images.append(im)
    criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)
 
    rows = 6 #number of checkerboard rows.
    columns = 9 #number of checkerboard columns.
 
    #coordinates of squares in the checkerboard world space
    objp = np.zeros((rows*columns,3), np.float32)
    objp[:,:2] = np.mgrid[0:rows,0:columns].T.reshape(-1,2)
 
    #frame dimensions. Frames should be the same size.
    width = images[0].shape[1]
    height = images[