Fix issues with "Edge Detection" example

[greg-kennedy]

There are two errors with this example.

• First, the kernel is incorrect. The center pixel should be weighted at 8, not 9.
• Second, Laplacian output should be centered about the medium-brightness pixel value (0.5), and clamped to the 0 to 1.0 range. The previous version began sum at 0, leading to negative values in the output... Another way of presenting this would be to keep sum at 0, but add a sum = abs(sum); at line 48.

Fixed the errors and added a few more comments at the top.

EDIT: I should add that the previous version wasn't necessarily "wrong" - it was definitely a "high-pass filter" i.e. finding changes of intensity, as it says in the description - but that is a bit different than what people commonly call "edge detection".

[greg-kennedy]

Someone should check the Javascript output and compare it to Java, it may actually be accurate now (and the comment about it being wrong could be removed)

[greg-kennedy]

On second thought, closing this for now and will open an issue to discuss.

[SableRaf]

From the original comment, it looks like the initial intention may have been to show how to sharpen an image rather than do literal Edge Detection.

This program analyzes every pixel in an image in relation to the neighboring pixels to sharpen the image.

For reference, here is the output for the original example:

And here is the output of the Laplacian Edge Detector as proposed by @greg-kennedy

I'm guessing the term "Edge Detection" may have been used loosely in the original example. How about changing the title of this example and making another pull request for a separate Edge Detection example?

[greg-kennedy]

Continuing the discussion at issue #11 and if the proposal seems OK, I'll submit a new PR with changes soon :)