使用opencv查找水平线和垂直线

Question

I am trying to pull the horizontal and vertical lines out of an image, to use them to find the points at which they intersect. I am using this page as a reference, but I think the problem is with the opencv object that I am not understanding. This reference is written in C++ but I am doing it in Python. When I display the horizontal image, I just have a completely black screen, although I should have an image of only the horizontal lines. Would anyone know what I am doing incorrectly?

Reference: https://docs.opencv.org/3.2.0/d1/dee/tutorial_moprh_lines_detection.html

Code:

import cv2
import numpy as np
img = cv2.imread('msss.png')
blur_img = cv2.GaussianBlur(img, (5, 5), 0)
gray_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
cv2.imshow('image', gray_img)
cv2.waitKey(0)
cv2.destroyAllWindows()
height, width = gray_img.shape
horizontal = gray_img.copy()
horizontal_size = int(width/30)
horizontal_struct = cv2.getStructuringElement(cv2.MORPH_RECT, 
                                             (horizontal_size, 1))
horizontal = cv2.erode(horizontal, horizontal, horizontal_struct, (-1, -1))
horizontal = cv2.dilate(horizontal, horizontal, horizontal_struct, (-1, -1))
cv2.imshow('image', horizontal)
cv2.waitKey(0)
cv2.destroyAllWindows()

Answer 1

Humm, seems that you are looking for how to identify corners. If yes you could try this reference already in python:

http://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_feature2d/py_shi_tomasi/py_shi_tomasi.html

Obs.: the link you provided is about morphological transformations. What they say "extract" is a reference for change the image such way those lines disappear without providing any reference of the lines coordinates. Usually these transformations are useful to make the image less noisily and help the next steps, like image segmentation, more easy.

Answer 2

You are looking for the hough transform. http://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_houghlines/py_houghlines.html