Hello everyone,
So i am trying to test a SLAM algo with OAK camera following the steps in the tuto of https://www.arducam.com/docs/opencv-ai-kit-oak/performing-location-with-visual-slam/#ftoc-heading-9.
I have a problem in the step 5.5 when building the last 3 targets kitti_gps_test, kitti_odom_test and vins_node. I have as an error :
usr/bin/ld : /home/mehdi/Bureau/stage/depthai/ros_workspace/OAK_Nvidia_Jetson_ROS_SLAM_VINS/VINS_GPU/devel/lib/libvins_lib.so : undefined reference to « cv::cuda::createGoodFeaturesToTrackDetector(int, int, double, double, int, bool, double) »
/usr/bin/ld : /home/mehdi/Bureau/stage/depthai/ros_workspace/OAK_Nvidia_Jetson_ROS_SLAM_VINS/VINS_GPU/devel/lib/libvins_lib.so : undefined reference to « cv::cuda::SparsePyrLKOpticalFlow::create(cv::Size_<int>, int, int, bool) »
I tried all solutions on internet but without any result.

  • erik replied to this.

    Hello Erik,
    Thank you for your reply, I will try this one instead.

    a year later

    Hello erik

    Might this totorial somehow will be implemented without LXD? Is LXD mandatory? Might anyone will posted here the step-by-step tutorial without LXD? Thanks.

    • erik replied to this.

      Hi ostepan2006 ,
      To be honest I am not sure - I would think if you're using ubuntu 20.04 (as what is specified in the article) there shouldn't be an issue if you are using all the same library versions.

        erik thanks.

        I am using ubuntu 18.04. Will be a problem with that?

        • erik replied to this.

          ostepan2006 My guess would be "yes" - and if you are asking me this (depthai support engineer, nothing with linux), you should very much follow the tutorial and use lxd.

            erik thanks

            As I undersatnd I have to have ubuntu 20.04 on host computer, will had run LXD-container and then I can will had implemented above mentonied tutorial. Right?