Python OpenCV HoughLinesP Inaccurate Line Detection

Question

Consider three points A, B, C in an image.

Below are their coordinates in the image of size 300x300.

I am trying to detect and draw a line connecting through these three points using below HoughLinesP code.

import cv2
import numpy as np

img = cv2.imread('test.png')
img = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY) #Convert img to grayscale 
lines = cv2.HoughLinesP(img, rho=1, theta=np.pi/180, threshold=1, minLineLength=5, maxLineGap=10)
print(lines)

for line in lines:
    x1, y1, x2, y2 = line[0]
    cv2.line(img, (x1, y1), (x2, y2), 255, 1)
cv2.imshow("result", img)

But it detects a line that passes through only B and C. Why is it so?

Output:
[[[110 100 120 100]]]

nathancy · Answer 1 · 2019-09-17T02:13:24.550

2

cv2.HoughLinesP() is mainly used for detecting a line, not really used for drawing. To draw a line given your three points, there are a few other options you can try. The first method is to filter the points by finding the leftmost and rightmost points then drawing the line with cv2.line(). Another approach is to find all the points then use cv2.fillPoly(). A third method is to use cv2.polylines()

import cv2
import numpy as np

image = cv2.imread('1.png')
gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)
coords = np.column_stack(np.where(gray > 0))
coords = np.rot90(coords, 2)
cv2.fillPoly(image, [coords], (36,255,12)) # or
# cv2.polylines(image, [coords], 1, (36,255,12))

cv2.imshow('image', image)
cv2.imwrite('image.png', image)
cv2.waitKey()

edited Sep 17 '19 at 02:13

answered Sep 17 '19 at 01:07

nathancy

42,661
14
115
137

HoughLinesP is mainly used for detecting lines, in your image you have three points instead of a line. Maybe after filling the points to create a line then you can use HoughLinesP – nathancy Sep 17 '19 at 20:28
Thank you. This dots and image is for just to understand the working of HoughLinesP with threshold parameter. And I believe it works by joining points (not only lines) that appears in straight line. – GreenHeart Sep 17 '19 at 20:44
HoughLinesP worked after increasing the Accumulator's resolution and threshold. – GreenHeart Sep 19 '19 at 13:58

GreenHeart · Accepted Answer · 2019-09-19T13:35:43.617

0

Increase the Accumulator's Resolution by giving much smaller values for rho and theta.
Set Accumulator's threshold to minimum 2.

lines = cv2.HoughLinesP(img, rho=0.1, theta=np.pi/180 * 0.1, threshold=2, minLineLength=5, maxLineGap=10)

edited Sep 19 '19 at 13:35

answered Sep 19 '19 at 13:19

GreenHeart

73
1
6

Python OpenCV HoughLinesP Inaccurate Line Detection

2 Answers2