2

I am using opencv and python .

I need to detect color of object in image for example given below image , the color of shirt is red.

enter image description here

enter image description here

On this link i found something useful but its detecting image of skin . http://lokeshdhakar.com/projects/color-thief/

I think i will have to use image contour extraction and then carry out color detection for that .

usernan
  • 572
  • 1
  • 6
  • 23

2 Answers2

6

Getting the dominant colors may be achieved using the following simple approach:

from sklearn.cluster import KMeans
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as patches

img = cv2.imread('red_shirt.jpg')
height, width, dim = img.shape

EDIT: take only the center of the image:

img = img[(height/4):(3*height/4), (width/4):(3*width/4), :]
height, width, dim = img.shape

img_vec = np.reshape(img, [height * width, dim] )

kmeans = KMeans(n_clusters=3)
kmeans.fit( img_vec )

EDIT: count cluster pixels, order clusters by cluster size

unique_l, counts_l = np.unique(kmeans.labels_, return_counts=True)
sort_ix = np.argsort(counts_l)
sort_ix = sort_ix[::-1]

fig = plt.figure()
ax = fig.add_subplot(111)
x_from = 0.05

for cluster_center in kmeans.cluster_centers_[sort_ix]:
    ax.add_patch(patches.Rectangle( (x_from, 0.05), 0.29, 0.9, alpha=None,
                                    facecolor='#%02x%02x%02x' % (cluster_center[2], cluster_center[1], cluster_center[0] ) ) )
    x_from = x_from + 0.31

plt.show()

enter image description here

You can remove BG and skin pixels with this kind of preprocessing

Community
  • 1
  • 1
Rosa Gronchi
  • 1,828
  • 15
  • 25
1
  1. Load the frame
  2. Convert BGR to HSV
  3. Range the value of pixel

also check out this link Color detection in opencv

Community
  • 1
  • 1
kini_dot
  • 33
  • 1
  • 6
  • 1
    Above sol which u said is when u want to detect given color example detect blue color in image . but i want find out what is color of image – usernan May 04 '16 at 09:46