-
Notifications
You must be signed in to change notification settings - Fork 2
dcintra/Computer-Vision_ObjectDetection
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
//--------------------------- // Computer Vision //--------------------------- System to detect objects based on object features. To match objects three criteria had to be met: 1. Matching # of holes 2. Aspect Ratio within +/-10% 3. Roundness Ratio within +/-10% I chose 800 rho bins and 180 such that I would end up with a range of -800:+800 rho values and -180:+180 degrees. Standard threshold for Hough: simply filtered anything that not within the 90th, 30th & 40th percentile for hough1/2/3 respectively. Line Segments: The key to my algorithm was to compute the line and then check whether all the points on the line had been detected by the edge detector. If a point was present in the edge detected image that meant that i'd include it as a point in my line. In a second step, once points were added to line segment matrices to be plotted; I would keep track of the distances between the points. If the distance between points passed a certain threshold, I'd start a new line segment.
About
Edge Detection, Template Matching using moments of inertia & Hough Transform
Resources
Stars
Watchers
Forks
Releases
No releases published