Edge Detection and Color Segmentation

Below are the results for the edge detection using the Sobel and Prewit method a star image.

Sobel Method

Prewitt Method

Next are the results for the edge detection using Sobel and Prewit method on a square image.

Sobel Method

Prewitt Method

Sobel and Prewitt methods that were applied to the star and square images results in images that look identical. No big difference can be visually observed from the edges detected from the two images using the two edge detection methods.

Applying the method to a real image with edges.

Applying the Sobel edge detection method to the image above.

Applying the Prewitt edge detection method to the image above.

After applying both Sobel and Prewitt edge detection methods on the image above. We can see that both of them actually performed sufficiently to outline the edges found in the image. This might be because the patterns of the edges present in the image are just fairly simple. The only thing that can be visually observed is that almost all of the edges are actually detected but some of them are not connected at their intersections.

Designing edge detection that works on diagonals and applying it to the image.

Applying this to the synthetic image.

Also applying this to the image.

After adding the diagonal filter, the main observation we can see is that the connections or intersections between the edges are now handled well than it was with the previous two methods.

Color Segmentation

An image of color pencils was used in this part. The image that was used can be seen below.

The light orange and dark orange colors were set as a mask and were segmented from the image, the result can be seen below.