Vanishing point detection

Implementation using Python and OpenCV library

Script vpdetector.py implements the method.

vpdetector.py takes up to 5 parameters (format: --parameterName value) and compulsory image path If a parameter is not submitted, default value is used.

Usage

python vpdetector.py -h
Detect vanishing points in jpg image.

positional arguments:
  img                   Image jpg file

optional arguments:
  -h, --help            show this help message and exit
  --cannyMin [CANNYMIN]
                        Canny edge detector: lower threshold [0, inf], default
                        100
  --cannyMax [CANNYMAX]
                        Canny edge detector: upper threshold [0, inf], default
                        300
  --houghTreshold [HOUGHTRESHOLD]
                        Hough t. threshold [0, inf], default 100
  --houghMinLineLen [HOUGHMINLINELEN]
                        Hough t. min. line length [0, inf], default 50
  --scoreThreshold [SCORETHRESHOLD]
                        Score threshold [0, 1], default 0.8

Example calls:

python vpdetector.py img "C:\Users\10250153\stack\vpoints\demo\5D4L1L1D_L.jpg" --cannyMin 80 --cannyMax 200 --houghTreshold 50 --houghMinLineLen 30 --scoreThreshold 0.8
python vpdetector.py img "C:\Users\10250153\stack\vpoints\demo\5D4L1L1D_L.jpg" --cannyMin 80 --houghTreshold 50 --houghMinLineLen 30 --scoreThreshold 0.75

All default parameter values

python vpdetector.py img "C:\Users\10250153\stack\vpoints\demo\5D4L1L1D_L.jpg"

Demos and data

Reseach materials

Algorithm overview

Intersection point weight computation