I really want to contribute and learn! But I can’t find any good first issues to start on, can someone please point me in a good direction or if they need help on an issue?
Short answer: the label exists but it isn’t maintained. label:"good first issue" matches zero open issues in opencv/opencv right now. It’s GitHub’s default label and no open issue currently uses it, so waiting for a tagged list won’t get you anywhere.
The official route is the wiki guide linked from CONTRIBUTING.md. Step 3 says: “Choose a task for yourself. It could be a bugfix or some new code.” That page points at the open label:bug query, which currently returns 916 issues. That’s your actual pool, not a curated beginner list.
Concrete next step:
- Pick a module you already use (imgproc, dnn, calib3d, videoio) and add its name to the bug filter, e.g.
is:issue is:open label:bug imgproc in:title. -
- Build the latest 4.x or 5.x from source. The guide requires confirming the bug still reproduces there; bugs that only affect deprecated 2.x/3.x are explicitly not fixed.
-
- Write a minimal repro first, then add a failing gtest to that module’s test file. Clone opencv_extra, set OPENCV_TEST_DATA_PATH, run ./bin/opencv_test_imgproc (swap in the test binary for your module).
-
- Open the PR against 4.x and reference the issue. CI runs through GitHub Actions and the buildbot, so you’ll see failures before a reviewer does.
If you want something smaller than an algorithm fix, docs and test-only changes go through the same branch/PR workflow.
- Open the PR against 4.x and reference the issue. CI runs through GitHub Actions and the buildbot, so you’ll see failures before a reviewer does.
One practical thing: leave a short comment on the issue when you start, so two people don’t write the same patch.
1 Like