You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Sep 15, 2023. It is now read-only.
Kyle Corry edited this page Jan 8, 2019
·
2 revisions
Contour filters are used to identify contours (solid shapes) in an image and decide which ones may look like a target and which are just noise.
Built-in filter types
StandardContourFilter: This is the recommended contour filter for most applications as it is the simplest.
ConvexHullContourFilter: The convex hull contour filter takes into consideration the perimeter, solidity, vertexes, area, and dimensions of contours.
Creating a StandardContourFilter object
area: The percent of the total area the target takes up. This helps filter out contours which are too small (noise) or too large (other objects).
fullness: The percent of the contour's area to the bounding box of its area. Values closer to 0 are contours which are not rectangular (such as 2016's vision target).
aspectRatio: The aspect ratio of the contours width / height. Giving a decent margin of error avoids having an actual target filtered out.