The Manifold 2G is a TX2 variant
Be default, it can grab image from DJI cameras by OSDK ROS 3.9.  

Then based on image, you can run darknet Yolo or other learning or non-learning model for actual running tasks.
by right you just need to follow this post to run. https://developer.dji.com/onboard-sdk/documentation/sample-doc/advanced-sensing-object-detection.html



But often you encounter problem such as

Failed call to cuInit: CUDA_EROR_NO_DEVICE


KerneQAM0QAM0QAM0QAM08HM18Nf45+GP8X//0MBeNbBnQ3HPuf0fX42kx3FcpWsAAAAASUVORK5CYII=[/img]
CUDA Error: __global__ function call is not configured:



In this case, you need to be in super user mode by just calling

sudo su

then
source ./catkin_ws/devel/setup.bash

or use other framework tests the resources, you should be able to run

hi sorry how did you manage to connect the manifold 2g to the dji m210 rtkndrone to access the camera? Do you have a step-by step tutorial on how to do it? I was planning to access the z30 camera for object deection. Do you have a picture of your hardware  connections? or specifically what do you write in the command because i do not understand anything in the website. It would be of great help if you can reply me asap thank you.