Circle Detection results

The following are the results of the Circle Detection at each stage in the processing pipeline

Original Image



This is a test image with 4 circles.

After Filtering Image

For the test data No filtering is required as the data set does not have any noise, hence the parameters were adjusted to give minimum blurring of the edges.
After Median Filtering (Window Size 3) After AnisoTropic Diffusion(Conductance 1.2, Iterations 3)

After Edge Detection

Canny Edge Detector (Threshold 10) Marr Hildreth Edge Detector (Gaussian 0.01)

Circle Detection Results

The Detected Circles

Original Image



This is an image with 4 coins.(3 quarters and a 1 penny)

After Filtering Image

After Median Filtering (Window Size 2) After AnisoTropic Diffusion(Conductance 1.2, Iterations 3)

After Edge Detection

The marr hildreth filter does not give good results for coin detection. Probably becoz the sphagetti forms spurious circles and makes detection of actual circles difficult.
Canny Edge Detector (Threshold 10) Marr Hildreth Edge Detector (Gaussian 0.01)

Circle Detection Results

The Detected Circles
Back To Main Page