使用 Python / Computer Vision 等获取图像中的顶点? (附图片)

Get vertices in a image using Python / Computer Vision , etc ? ( Images Attached )

你好,我需要在图像中找到给定形状的顶点(x 和 y 坐标),在进行分割和边缘提取之后,以下是获得的图像:

以下是我需要找到其坐标的顶点:

我想你可能想先用霍夫线变换来找到线条。然后,您可以从检测到的线中获取交点。你可能会找到OpenCV关于霍夫线变换的教程here

这是我使用霍夫线变换的结果:

代码:

import numpy as np
import cv2 as cv2
import math

img_path = 'hSAdf.png'

# Read the original image
img = cv2.imread(img_path) 

# Convert to graycsale
img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
dst = cv2.threshold(img_gray, 50, 255, cv2.THRESH_BINARY)[1]

cdst = cv2.cvtColor(dst, cv2.COLOR_GRAY2BGR)

lines = cv2.HoughLines(dst, 1, np.pi / 180, 180, None, 0, 0)

# Drawing the lines
if lines is not None:
    for i in range(0, len(lines)):
        rho = lines[i][0][0]
        theta = lines[i][0][1]
        a = math.cos(theta)
        b = math.sin(theta)
        x0 = a * rho
        y0 = b * rho
        pt1 = (int(x0 + 10000*(-b)), int(y0 + 10000*(a)))
        pt2 = (int(x0 - 10000*(-b)), int(y0 - 10000*(a)))
        cv2.line(cdst, pt1, pt2, (0,0,255), 3, cv2.LINE_AA)

cv2.imshow("Detected Lines (in red) - Standard Hough Line Transform", cdst)
cv2.imwrite("output.png", cdst)
cv2.waitKey(0)

这里我没有使用 Canny 边缘检测,因为我认为图像本身非常 line-clear,这使得边缘检测变得多余。

函数HoughLines() returns直线的rho(像素)和theta(弧度),对应直线方程:

编辑 1:rho、theta 和 m、c 之间的简单转换:

m = tan(theta + PI/2)

c = rho / sin(theta)

图片来自Socret Lee

我想你可以继续调整线路检测功能。您可以手动调整阈值,甚至可以在函数中限制线​​条的梯度。然后,您可以通过裁剪和限制渐变来定位一行。

或者您可以拒绝相差约 90 度的线的交点。然后,您将获得所需的积分。

使用Contour detection和近似你可以得到外部顶点,并计算它们:

[1737 197] [616 199] [225 596] [ 226 1708] [ 610 2102] [1717 2121] [2118 1732] [2134 601]

#!/usr/bin/python3
# -*- coding: utf-8 -*-

import cv2
import numpy as np

img = cv2.imread("input.png", 0)

def fillhole(input_image):
    '''
    input gray binary image  get the filled image by floodfill method
    Note: only holes surrounded in the connected regions will be filled.
    :param input_image:
    :return:
    '''
    im_flood_fill = input_image.copy()
    h, w = input_image.shape[:2]
    mask = np.zeros((h + 2, w + 2), np.uint8)
    im_flood_fill = im_flood_fill.astype("uint8")
    cv2.floodFill(im_flood_fill, mask, (0, 0), 255)
    im_flood_fill_inv = cv2.bitwise_not(im_flood_fill)
    img_out = input_image | im_flood_fill_inv
    return img_out 

res = fillhole(img)

contours = cv2.findContours(res, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]

peri = cv2.arcLength(contours[945], True)
approx = cv2.approxPolyDP(contours[945], 0.04 * peri, True)

im = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
s = 10
for p in approx:
    p = p[0]
    print(p)
    im[p[1]-s:p[1]+s, p[0]-s:p[0]+s] = (255, 255, 0)

cv2.drawContours(im, contours, 945, (0, 200, 255), 3)

cv2.namedWindow("img", cv2.WINDOW_NORMAL)
cv2.imshow("img", im)
cv2.waitKey(0)