How can I improve the results of HoughLinesP() in my OpenCV Python script

Question

I am trying to get all the lines in this image:

This is the code that I'm using:

threshold = 30
minLineLength =10
maxLineGap = 10
lines = cv2.HoughLinesP(img,1,np.pi/360, threshold, minLineLength, maxLineGap)

The problem is that I'm getting too many lines (~300):

But if I increase the threshold value it starts to miss some lines:

Is there any way of reducing the number of lines while keeping line-detection accurate?

Thanks in advance!

Answer 1

It works (mostly) fine for me in Python/OpenCV. Adjust your HoughP line arguments as appropriate.

I think you need to threshold your image first. And perhaps thin the white lines.

Input:

import cv2
import numpy as np

# read image as color not grayscale
img = cv2.imread("lines.png", cv2.IMREAD_COLOR)

# convert img to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# do threshold
thresh = cv2.threshold(gray, 30, 255, cv2.THRESH_BINARY)[1]

# get hough line segments
threshold = 30
minLineLength =10
maxLineGap = 10
lines = cv2.HoughLinesP(thresh, 1, np.pi/360, threshold, minLineLength, maxLineGap)

# draw lines
results = img.copy()
for [line] in lines:
    print(line)
    x1 = line[0]
    y1 = line[1]
    x2 = line[2]
    y2 = line[3]
    cv2.line(results, (x1,y1), (x2,y2), (0,0,255), 1) 

# show lines
cv2.imshow("lines", results)
cv2.waitKey(0)

# write results
cv2.imwrite("lines_hough.png",results)

Resulting Hough lines in red:

You get a lot of parallel very close lines that you may want to merge somehow or thin out the list.