基于python opencv实现广角相机标定和图像畸变矫正

最新推荐文章于 2022-10-30 14:07:44 发布

studyer_domi

最新推荐文章于 2022-10-30 14:07:44 发布

阅读量2.4k

点赞数

分类专栏： python 图像处理 opencv 文章标签： python opencv 相机标定畸变

本文链接：https://blog.csdn.net/qingfengxd1/article/details/108539704

版权

python 同时被 3 个专栏收录

126 篇文章 93 订阅

订阅专栏

图像处理

42 篇文章 5 订阅

订阅专栏

opencv

20 篇文章 1 订阅

订阅专栏

目的：

实现相机标定，得到相机的内参以及畸变旋转参数等
尝试矫正由相机产生的图像畸变

代码：

import cv2 as cv
import numpy as np
import glob
import os

#循环中断
criteria=(cv.TERM_CRITERIA_EPS+cv.TERM_CRITERIA_MAX_ITER,30,0.001)

#标定板交叉点的个数
row=6
column=4
objp=np.zeros((row*column,3),np.float32)
objp[:,:2]=np.mgrid[0:row,0:column].T.reshape(-1,2)

objpoints=[]   #实际空间3D点
imgpoints=[]   #图像中2D点

#批量读取图片
images=glob.glob('D:\\python\\photos\\chess\\*.jpg') 

for fname in images:
    img=cv.imread(fname)
    gray=cv.cvtColor(img,cv.COLOR_BGR2GRAY)
    
    #找标定板角点
    ret,corners=cv.findChessboardCorners(gray,(row,column),None)
    
    if ret==True:
        objpoints.append(objp)
        
        corners2=cv.cornerSubPix(gray, corners, (11,11), (-1,-1), criteria)
        imgpoints.append(corners2)
        
#标定相机
ret,Matrix,dist,rvecs,tvecs=cv.calibrateCamera(objpoints,imgpoints,gray.shape[::-1],None,None)

datadir="D:\\python\\photos\\chess\\"

path=os.path.join(datadir)
img_list=os.listdir(path)

for i in img_list:
    img=cv.imread(os.path.join(path,i))
    h,w=img.shape[:2]                                    
    newMatrix, roi = cv.getOptimalNewCameraMatrix(Matrix, dist, (w,h), 1, (w,h)) #矫正图像

    dst = cv.undistort(img, Matrix, dist, None, newMatrix)
    cv.imwrite('D:\\python\\practice\\photos\\calibrated\\'+i,dst)

#计算重投影误差
tot_error = 0
for i in range(len(objpoints)):
    imgpoints2, _ = cv.projectPoints(objpoints[i], rvecs[i], tvecs[i], Matrix, dist)
    error = cv.norm(imgpoints[i],imgpoints2, cv.NORM_L2)/len(imgpoints2)
    tot_error += error

#输出参数
print('ret:\n',ret)
print('mtx:\n',Matrix)
print('dist:\n',dist)
print('rvecs:\n',rvecs)
print('tvecs:\n',tvecs)
print ("total error: ", tot_error/len(objpoints))

欢迎关注公众号：算法工程师的学习日志