Closed benbelov closed 5 years ago
Rip I'll get on it during the meeting today
On Sun, Dec 2, 2018, 10:22 AM Tianshu Huang <notifications@github.com wrote:
@thetianshuhuang requested changes on this pull request.
Let's make a few structural changes to the testing interface (I'll make DBScan retrofit their code).
In vision/convexhull/python/convexhull_test.py https://github.com/ut-ras/r5-2019/pull/34#discussion_r238104814:
+ +def pairwise(iterable):
- "s -> (s0,s1), (s1,s2), (s2, s3), ..."
- a, b = tee(iterable)
- next(b, None)
- return zip(a, b)
+for a, b in pairwise(result):
- cv2.line(img,a,b,(0,0,255),5)
+#print(result) +cv2.namedWindow("corners", cv2.WINDOW_NORMAL) +cv2.resizeWindow("corners", 120, 120) +cv2.imshow("corners", img) +cv2.waitKey(0) +cv2.destroyAllWindows()
Change the test to be a callable function that takes the file as an argument, and saves the output as a file.
Let's start a standard format for vision tests:
- Call the tests folder tests (vision/convexhull/tests)
- Make each test
; ex - convexhull_640_480.png (no metadata on your images) - Instead of using the directory relative to the file ( ../tests/convexhull_640_480.png), use paths relative to the file:
BASE_DIR = os.path.join(os.dirname(file), "tests") img = cv2.imread(os.path.join(BASE_DIR, "convexhull_640_480.png"))...
I'll ask DBScan to make these changes to their branch as well.
— You are receiving this because your review was requested. Reply to this email directly, view it on GitHub https://github.com/ut-ras/r5-2019/pull/34#pullrequestreview-180567845, or mute the thread https://github.com/notifications/unsubscribe-auth/AeD8pGJrDnrsF7l6RavJjndRNZwuJZTcks5u0_4-gaJpZM4Y9Snu .
Description
Added convexhull vision module in Python.
API Changes
Addition of
convexhull
function.Todos