How to detect the traffic circle / ring?

[Lamp]

hello,

how can I detect the traffic circle / ring on the image?

please guide the API to be used.

ring:

0 0 0 0 0 0
0 1 1 1 0 0
1 1 0 1 1 0
0 1 0 1 0 0
0 0 1 0 0 0

Regards

[Naveen_G_Intel]

Hi xiehuan,

i am not sure about traffic circle / ring on the image means, Sorry. Could you explain little more?

Thanks,

Naveen Gv

[Lamp]

Hi Naveen,

Thanks for your reply.

I didn't describeit clearly, sorry, we want to recognize the black area on the below image:






and we want to exclude the black area that is not ring:





I hope the image can be displayedproperly.



Regards,
Xie Huan

[Lamp]

I'm not sure if the image can be viewed or not, so I'll try to explain it by word.

Imaging there is an image with white background (pixel value = 1), and there is a blackcircle (pixel value = 0)on the center of the image; another circle with white color (pixel value = 1) has the same center as the previous one, but the radius is less than the previous one, covers the black circle.

we want to recognize the black area, which is a part of the black circle but not all.
all the circles that contains a hole can be consider as a ring / traffic circle, andexclude thesolid circle.

Regards,
Xie Huan

[Vladimir_Dudnik]

Hello,

it seems you are looking for circle detection algorithm. According many sources the circular Hough transform might be used to detect circles of known radius. In IPP we do have Hough transform implementation to detect lines but do not have one for circles. You may take a look at OpenCV library which provides circular hough transform implementation.

Regards,
Vladimir

[Naveen_G_Intel · Accepted solution]

Hi Xie Huan,

As Vladimir said, OpenCV has got functions which finds circles in a grayscale image using a Hough transform.

If you want to see Hough transform functions available in IPP click here, but this can be used only for detecton of the straight lines that are defined by the parametric equation

r = x*cos() + y*sin(), where r and are the length and angle from the origin of a normal to the line respectively

Regards,

Naveen Gv