Sobel Filter¶

This is a filtering method used to identify and highlight coarse changes in pixel intensity based on the 1st derivative.

sobel_filter(img, k, scale, device, debug=False)

returns device, filtered image

• Parameters:
  • img - binary image object. This image will be returned after filling.
  • dx - derivative of x to analyze (1-3)
  • dy = derivative of y to analyze (1-3)
  • k - apertures size used to calculate 2nd derivative filter, specifies the size of the kernel (must be an odd integer: 1,3,5...)
  • scale - scaling factor applied (multiplied) to computed Laplacian values (scale = 1 is unscaled)
  • device - Counter for image processing steps
  • debug- Default value is False, if True, filled intermediate image will be printed
• Context:
  • Used to define edges within and around objects
• Example use:
  • Use In NIR Tutorial

Original grayscale image

import plantcv as pcv

# Apply to a grayscale image
# Filtered image will highlight areas of coarse pixel intensity change based on 1st derivative
device, lp_img = pcv.sobel_filter(img, 1, 0, 1, device, debug=True)
device, lp_img = pcv.sobel_filter(img, 0, 1, 1, device, debug=True)

Sobel filtered (x-axis)

Sobel filtered (y-axis)