8000 GitHub - felipeasimos/reconstructor
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

felipeasimos/reconstructor

Repository files navigation

3D Reconstruction of an image (Manhattan-like)

How to run:

pip3 install -r requirements.txt
python3 main.py IMAGE_PATH

1. Edge Detection

Whare are image gradients?

It is a directional change in the intensity or color in an image. In the edge detection context, borders show a sudden change in these attributes between close pixels. Intuition: imagine the image as a heightmap, gradients then map directly to the idea of derivatives as the inclination of a 2d function

Roadmap

  1. Grayscale
  2. Canny - in the output every edge has a 1 pixel thickness
    1. Gaussian blur
      • Reduces noise
    2. Compute gradients magnitude and direction using sobel
      • Two Sobel filters compute change in the x and y directions
      • Combined we can get the actual direction of change (gradient direction and magnitude)
    3. Non-maximum suppresion
    4. Double thresholding and edge tracking
  3. Hough

2. Vanishing Points

  1. Get hough line intersection points
  2. Cluster intersection points to get the dominant vanishing points
  3. Find 3 orthogonal vanishing points

3. Calibration

  1. Calculate omega (absolute conic)
  2. Calculate K

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  

Languages

0