Solve the problem of the test item test_multi_camera_aligment
A previous problem with this test project was that circle in the
ultra-captured images were not recognized. The analysis found that
because the FOV of ultra is too large, the circle in the image are
too small, and it will be filtered out. (The script will filter out
contours with smaller areas)
The current solution is to relax the filtering conditions and make
the minarea smaller.
Bug: 214881385
Signed-off-by: kongchenrui <kongchenrui@xiaomi.com>
Change-Id: Ie37a024a6f7f373fd127c21838a1e4fdc0093e4f
(cherry picked from commit 811c621630d7fd9019feb3c2132d9f5487db9bbb)
diff --git a/apps/CameraITS/tests/scene4/test_multi_camera_alignment.py b/apps/CameraITS/tests/scene4/test_multi_camera_alignment.py
index 73d1276..af55ebc 100644
--- a/apps/CameraITS/tests/scene4/test_multi_camera_alignment.py
+++ b/apps/CameraITS/tests/scene4/test_multi_camera_alignment.py
@@ -29,7 +29,7 @@
import opencv_processing_utils
ALIGN_TOL_MM = 4.0 # mm
-ALIGN_TOL = 0.01 # multiplied by sensor diagonal to convert to pixels
+ALIGN_TOL = 0.0075 # multiplied by sensor diagonal to convert to pixels
CIRCLE_COLOR = 0 # [0: black, 255: white]
CIRCLE_MIN_AREA = 0.01 # multiplied by image size
CIRCLE_RTOL = 0.1 # 10%