Image Processing - Color Segmentation

This is a simple application with OpenCV Library using HSV color system to segment the blue spectrum color on a given input image, in this case, video frames from a webcam.

the procedure was:

1 - Convert the RGB image to HSV
2 - Split HSV image and get the Hue Plane
3 - make all values outside 90~135 range as zero (H S and V values), remember that opencv divide the real value of an HSV image by 2
4 - smooth the result image using a gaussian filter 15x15
5 - convert back the result image to RGB color system
6 - make it grayscale

you can use mathematical morphology to improve the results (dilate and erode functions) and then binarize it.

Category:

Science & Technology

Tags:

• opencv
• color detection
• image processing
• computer vision
• processamento de imagem
• deteccao de cor
• visao computacional
• segmentation
• segmentacao

[femistofel]

after converting to hsv you could take just the hue plane and work with that. Seems to me it would save up some processing power.

[DDoiS]

Yes, I'm doing that, I just forgot to put it on description field. Thanks =D

[roseworld11]

hi brother can I get ur email, cuz I have project n I faced some problems

[doliharahap]

hai bro..

do you know how if i want to work just with the V value plane??

do you have any idea??