ModalAI Forum

@Caio-Licínio-Vasconcelos-Pantarotto the VIO algorithm takes the starting extrinsics and refines based on movement of the system. So, it's normal to see small updates to the value

@Caio-Licínio-Vasconcelos-Pantarotto @Moderator I had a similar question, I calibrated the TOF camera using ros cv calibration available online. I wanted to know is there a way possible to generate a aligned depth image, since the TOF and RGB sensors have different resolutions. Here are the two questions which I asked about the same issue: https://forum.modalai.com/post/12529 @Darshit-Desai said in VOXL MPA_to_ROS rostopics not subscribing to anything: @Darshit-Desai said in VOXL MPA_to_ROS rostopics not subscribing to anything: Is there a way to get camera intrinsic Matrix of the ToF camera. Also, if I want to align the depth image with the RGB image how do I do that? Hi sorry, I wanted to ask about the intrinsics that was highlighted by mistake. @ben-grocholsky https://forum.modalai.com/post/12472 @Darshit-Desai said in Voxl2 Camera calibration fails for hires camera: @Moderator Ok I will try calibrating with the ros method of calibration. I was asking this because I want to make an aligned depth image by combining hires and tof camera like the realsense cameras have. If you could also reply on the following. I am not really sure if Tofs can be calibrated @Darshit-Desai said in Voxl2 Camera calibration fails for hires camera: Another Requests: How to do calibration of Time of flight camera or is there a intrinsic matrix?

@Jetson-Nano Please do not attempt to update the firmware with STLink. We will send you new boards properly programmed.

@Eric-Katzfey Thank You! it worked great.

@Jskim Can you please open these new questions as new topics? That way the appropriate staff member can address them. Thanks!

@James-Strawson Thanks for your reply! It made me realise what the error must be. We use a locally cloned repository for libmodalpipe because we want to build our software directly on the drone. This is then used in the build and link process of our software. This happened to be on the 'master' branch and not the 'SDK-1.0.0' branch. So the pipe clients use a different version than the pipe servers which is likely the cause of our errors.

@Chad-Sweet thank you

Each service is using MPA, so there are many pipes going at any given time. With two TOF, you are likely overheating the CPU Do you have airflow on the board? Are you using voxl-inspect-cpu to monitor how hot you are getting the CPU?

Hi @Ege-Yüceel, We usually create a volume to those headers when doing something like this but you can install the libmodal pipe package inside the docker container like this if you want. wget http://voxl-packages.modalai.com/dists/qrb5165/sdk-1.5/binary-arm64/libmodal-pipe_2.13.1_arm64.deb dpkg -i libmodal-pipe_2.13.1_arm64.deb **Note change sdk-1.5 to whatever sdk version you're on

@marian Oh yes, thanks for letting us know. When you are using raw frames over USB 2.0 you have to go to a much lower frame rate or else there isn't enough bandwidth to transfer them.

Hi @Nishkala , With the warning that I've not yet tried in python but I'm sure it will work This is the source code for our voxl-inspect-qvio tool, and from there we can see it's looking for a pipe QVIO_EXTENDED_LOCATION https://gitlab.com/voxl-public/voxl-sdk/services/voxl-qvio-server/-/blob/master/clients/voxl-inspect-qvio.c#L51 That's defined here: https://gitlab.com/voxl-public/voxl-sdk/services/voxl-qvio-server/-/blob/master/common/voxl_qvio_server.h#L64 The other place to look for defines is in libmodal-pipe: https://gitlab.com/voxl-public/voxl-sdk/core-libs/libmodal-pipe/-/blob/master/library/include This is where the inspect tool consumes data: https://gitlab.com/voxl-public/voxl-sdk/services/voxl-qvio-server/-/blob/master/clients/voxl-inspect-qvio.c#L259 Here's where we parse and print the info: https://gitlab.com/voxl-public/voxl-sdk/services/voxl-qvio-server/-/blob/master/clients/voxl-inspect-qvio.c#L184

Glad to hear someone's building our packages themselves! Yes, we're most of the way through migrating to a more platform independent build system that will let us use these packages on the qrb5165 platform as well. This is one of the things we've had to change, the next voxl system image will have a tweak allowing these packages, but if you want to install them in the meantime, you can manually add the tweak by adding the line arch arm64 7 to /etc/opkg/arch.conf on your voxl and running opkg update

https://gitlab.com/voxl-public/voxl-sdk/services/voxl-portal/-/blob/master/src/video_manager.cpp#L77 Is an example of allocating a buffer and copying the frame into it. There are some other good examples in that file as well for processing frames and converting them to various formats within the callback. Hope that helps!

@Eric-Katzfey Ok thanks for the fast answer.

@Matt-Turi Thank you

Hey @Michael-Shomin, The non-cached files in voxl-portal for font + images were purged in these two commits: https://gitlab.com/voxl-public/modal-pipe-architecture/voxl-portal/-/commit/d79786562275bcc11a41a1ebdf8c4c4de6ce5432 https://gitlab.com/voxl-public/modal-pipe-architecture/voxl-portal/-/commit/96259ed613f336ea4f2ac88ecd7a79918d8494a1 These changes are in the dev release, but you can pull them into your version if doing anything custom.

@Philemon-Benner Our qrb5165 based platforms put 64 bit libraries into /usr/lib. On Voxl the 64 bit libraries go into /usr/lib64. So the 64 bit support is specifically for our qrb5165 based platforms (e.g. RB5 Flight).

Thanks for the answer! it worked great after the upgrades. what I am trying to do now is to run the deeplab dnn, but I am having some issues. I tried to run the 865 build but i couldn't run any tflite dnn. then I tried to run the 820 build which worked perfectly and than added manually the deeplab dnn. eventually i received the error: [image: 1646902913214-screenshot-from-2022-03-10-10-55-20.png] do you have any suggestions?

@Philemon-Benner You can only connect one client at a time. But you can run multiple instances of voxl-streamer and give them different ports for clients to connect to. However, the best way to allow multiple clients is to setup a media server of some sort (eg Live555) so that you only have one stream going over the wireless link to the media server and then it can replicate it as needed over a wired network.