How to modify pixel values in numpy array of HSV image data?

Question

I have numpy array of HSV image data:

Shape:

(960, 1280, 3)

Data:

 [[ 90  53  29]
  [ 90  53  29]
  [ 68  35  29]
  ..., 
  [ 66  28 146]
  [ 58  21 145]
  [ 58  21 145]]

 [[ 90  53  29]
  [ 90  53  29]
  [ 75  35  29]
  ..., 
  [ 65  31 148]
  [ 69  18 144]
  [ 69  18 144]]]

I want to create a filter (eg "H < 20 or V > 200") and based on that filter to modify the array so that I can set HSV values to something I need, like [0 0 0].

I couldn't wrap my head around the indexing system, how this should be approached?

Hannes Ovrén · Accepted Answer · 2015-09-04 11:29:52Z

1

You create a mask array to pick the elements you want to change:

H = image[:,:,0]
V = image[:,:,2]
mask = (H < 20) & (V > 200)
image[mask] = 0

answered Sep 4, 2015 at 11:29

Hannes Ovrén

22k10 gold badges71 silver badges76 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Harriv Over a year ago

Thanks, I "knew" it must be simple.. :) What if I want to define something else as a value, eg [127 64 100]?

Divakar · Accepted Answer · 2015-09-04 12:06:45Z

Assuming hsv as the input hsv image data, you can use some reshaping alongwith boolean indexing to set all three channels to a generic triplet like so -

newvals = np.array([127,64,100])

mask = (hsv[:,:,0] < 20) | (hsv[:,:,2]> 200)
hsv.reshape(-1,3)[mask.ravel()] = newvals

Sample run -

In [13]: hsv
Out[13]: 
array([[[155, 179, 207],
        [200,  52, 185],
        [241, 139, 232],
        [188, 149, 117]],

       [[145, 169, 116],
        [146, 134, 108],
        [ 74,  34, 121],
        [  9, 190,  91]],

       [[240, 207, 228],
        [140, 158, 124],
        [179, 154, 212],
        [ 79, 166, 131]]], dtype=uint8)

In [14]: newvals = np.array([127,64,100])
    ...: mask = (hsv[:,:,0] < 20) | (hsv[:,:,2]> 200)
    ...: hsv.reshape(-1,3)[mask.ravel()] = newvals

In [15]: hsv
Out[15]: 
array([[[127,  64, 100],
        [200,  52, 185],
        [127,  64, 100],
        [188, 149, 117]],

       [[145, 169, 116],
        [146, 134, 108],
        [ 74,  34, 121],
        [127,  64, 100]],

       [[127,  64, 100],
        [140, 158, 124],
        [127,  64, 100],
        [ 79, 166, 131]]], dtype=uint8)

Collectives™ on Stack Overflow

How to modify pixel values in numpy array of HSV image data?

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related