yuxingyun / pylsd2

python bindings for Line Segment Detection algorithm (LSD and EDLines)

pylsd2

1. Introduction

pylsd2 is the python bindings for Line Segment Detection algorithm, including LSD and EDLines.

• LSD part is forked frompylsd, and upgrade lsd from 1.5 to 1.6, thanks primetang
• EDLines part is forked fromline_detector, thanks frotms

Windows and linux is supported currently, merge request for mac is welcome

2. Install

directly through pip to install it:

[sudo] pip install pylsd2

3. Usage

• by using cv2 module

import cv2
from pylsd2 import LineSegmentDetection, LineSegmentDetectionED
fullName, out_path = 'car.jpg', 'car_lsd.jpg'
src = cv2.imread(fullName, cv2.IMREAD_COLOR)
gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
# input single channel image
lines = LineSegmentDetection(gray)  # [(x1, y1, x2, y2, width, p, log_nfa), ...]
# lines = LineSegmentDetectionED(gray)  # [(x1, y1, x2, y2), ...]
for l in lines:
    pt1, pt2 = tuple(l[:2]), tuple(l[2:4])
    cv2.line(src, pt1, pt2, (0, 0, 255), 1)
cv2.imwrite(out_path, src)

• by using PIL(Image) module

from PIL import Image, ImageDraw
import numpy as np
from pylsd2 import LineSegmentDetection, LineSegmentDetectionED
fullName, out_path = 'car.jpg', 'car_edlines.jpg'
img = Image.open(fullName)
gray = np.asarray(img.convert('L'))
# input single channel image
# lines = LineSegmentDetection(gray)
lines = LineSegmentDetectionED(gray)
draw = ImageDraw.Draw(img)
for l in lines:
    pt1, pt2 = l[:2], l[2:4]
    draw.line((pt1, pt2), fill=(0, 0, 255), width=1)
img.save(out_path)

more examples:

The following is the result:

• car.jpg Original

  

• with lsd algorithm

  

• with EDLines algorithm

  

• house.png Original

  

• with lsd algorithm

  

• with EDLines algorithm

  

• chairs.jpg Original

  

• with lsd algorithm

  

• with EDLines algorithm

  

4. Compile Library from cpp

4.1 Compile For LSD

** Windows **

cd source/lsd && mkdir build && cd build
cmake ..

open the LSD.sln with Visual Studio, and Build The "ALL_BUILD" Project ** Linux **

cd source/lsd && mkdir build_so && cd build_so
cmake ..
cmake --build .

4.2 Compile For EDLines

** Windows **

cd source/edlines && mkdir build && cd build
cmake ..

open the EDLines.sln with Visual Studio, and Build The "ALL_BUILD" Project ** Linux **

cd source/edlines && mkdir build_so && cd build_so
cmake ..
cmake --build .