Timeline for Improving Canny Edge Detection and Contours Image Segmentation
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 23, 2017 at 12:41 | history | edited | CommunityBot |
replaced http://stackoverflow.com/ with https://stackoverflow.com/
|
|
| Feb 4, 2017 at 21:06 | comment | added | errcw | I'm using OpenCV for the Hough transform so I expect that aspect is correct. What I am less certain of is how to best use the result. For my example images I find the edge detection is either very noisy and produces too many lines to be useful, or too strict and fails to identify critical card edges. Moreover, even when all the necessary edges are present, I failed to find Hough parameters that consistently identified all the card edge lines (especially in the presence of artifacts like glare obscuring edges). | |
| Feb 4, 2017 at 15:28 | comment | added | Maximilian Matthé | @errcw there are various tested implementations for the hough-Transform for lines. Maybe, you can double-check your implementation against these. | |
| Feb 4, 2017 at 2:49 | comment | added | errcw | Thanks for the pointers! I've spent the last few days working through an implementation of the Hough transform for my use case. Unfortunately I've discovered that the edge detection appears to be too noisy to produce particularly good results (that, or my implementation is flawed, which is perhaps more likely). I'll keep poking at it because conceptually it seems like a powerful & relevant approach, but given my lack of success so far I'm happy to hear other options too. | |
| Jan 30, 2017 at 7:50 | history | answered | Maximilian Matthé | CC BY-SA 3.0 |