Talk:Generalised Hough transform

quality of description

I find this article awful. Let's ignore that it's hugely paraphrased from the Ballard paper. The Ballard paper itself isn't all that illuminating on details and intuition.

It needs concrete examples, i.e. illustrations various situations, concrete values instead of "0, Δɸ, 2Δɸ" (copied from the paper) which is absolute nonsense since Δ is never introduced and ɸ is supposed to be variable, not a constant. There's also no explanation of the multiple R entries in the third column of the table (I see why that can be the case but it better be explained with an example!). Also, why is position denoted by x and y, when those are customarily the x and y coordinates of a point? Another consequence of copying the paper rather than adding clarity. Use different variables to denote points, say p and q. Further, a point is two degrees of freedom and that should be made crystal clear in "a={y,s,θ}", which should also not be a set but a tuple!

Also there's a "bug". Detection (0) should talk about the haystack image, not the needle image.

I'm afraid to put any effort into rewriting this article because wikipedia is known for edit wars and individuals defending their favorite articles against any improvements. If someone with "clout" would give me a sign of potential support in the bureaucracy/politics aspect of this, you may be getting some contributions. --78.35.195.71 (talk) 23:35, 5 December 2020 (UTC)