search

halajun / VDO_SLAM

VDO-SLAM: A Visual Dynamic Object-aware SLAM System
Other
746 stars 115 forks source link

Problem when running demo-omd #38

Open yxianzhe opened 2 years ago

yxianzhe commented 2 years ago

Hi, this work is really fantastic. When I tried demo-kitti, it worked well and got a very good estimation. I just had some problem when running demo-omd. And I had the following mistake:

..........Dealing with Camera Pose.......... terminate called after throwing an instance of 'cv::Exception' what(): OpenCV(4.2.0) /home/yxz/lib_sources/opencv/opencv/modules/calib3d/src/solvepnp.cpp:216: error: (-215:Assertion failed) npoints >= 4 && npoints == std::max(ipoints.checkVector(2, CV_32F), ipoints.checkVector(2, CV_64F)) in function 'solvePnPRansac'

I tried to modify the variable reprojectionError to have more good matches, but it didn't work. Do you have any ideas about this mistake?

ZhenhuiL1n commented 2 years ago

I also have this problem.

RGthesis commented 2 years ago

Hi, I am a begginer in this project and i am facing the same issue as you @yxianzhe @ZhenhuiL1n . If you have found a solution, let me know ;)

JonnyG123789 commented 1 year ago

same

JonnyG123789 commented 1 year ago

it seems like too few matched keypoints

JonnyG123789 commented 1 year ago

Hi, I am a begginer in this project and i am facing the same issue as you @yxianzhe @ZhenhuiL1n . If you have found a solution, let me know ;)

DId you got this problem?

JonnyG123789 commented 1 year ago

The method is to "#" about depth value in Tracking.cc 200.

Sherryblues commented 1 year ago

The method is to "#" about depth value in Tracking.cc 200. hello! I found that line 200 in Tracking.cc has already been annotated. Can you explain it in more detail？ Thank you! 图片

sinsyamoses commented 1 year ago

i also have this problem

hu-xue commented 1 year ago

The method is to "#" about depth value in Tracking.cc 200. hello! I found that line 200 in Tracking.cc has already been annotated. Can you explain it in more detail？ Thank you! 图片

if we use omd dataset, it is based on the RGB-D camera. So we should change the depth calculate function of RGB-D camera, such as line 193. so uncomment 193, and comment 191 can fix the problem.