I have a project where the D435 camera will detect objects and make the drone stop at some distance before the object. Recognition of object is not needed. Is there any examples where I can have reference to and follow?

[MartyG]

Intel sell drone equipment that can make use of RealSense, known as the 'Intel Aero' brand. An advantage of an Aero-based approach is that you can make use of flight controller software with obstacle avoidance, such as 'DroneCode PX4'.

DroneCode Px4's instructions have been updated to include the current 400 Series camera models (D415 and D435)

https://docs.px4.io/v1.9.0/en/complete_vehicles/intel_aero.html

These instructions are based around use of the software on a single-board mini computer called the Intel Aero Compute Board.

EDIT: Intel Aero is now a retired brand and the Intel store will not be re-stocked with hardware parts for it.

https://click.intel.com/intel-aero-platform-for-uavs-compute-board-2772.html

However, use of Intel Aero parts is not an absolute requirement to get D435 working on a drone if you prefer to use the parts that you already have. The link below discusses a user who has such a setup with obstacle avoidance on a drone. Looking at the launch file they supplied, they were using ROS with their drone in ROS' RGBD mode.

https://forums.intel.com/s/question/0D50P0000490XM0SAM/noise-in-realsense-d435-pointcloud

 Their ROS launch file and an MP4 short video file of a flight with the D435 are attached to their message. You can download the video by left-clicking on the attached MP4.

[EYeo]

Thank you for your response. We already have the realsense viewer in Odroid and the camera running. Is it possible to extract the raw data in term of numbers in order to proceed with the programming.

[MartyG]

If you just want to detect the depth of an object on the image and get a value for its distance from the camera, the RealSense SDK documentation has an example script on its front page.

https://github.com/IntelRealSense/librealsense#ready-to-hack

It would be possible to expand upon that script to do an 'If' logic statement that checks the distance value and IF it is a certain distance or less from the camera, trigger the drone to stop.

[EYeo]

How do I install the pyrealsense2 package onto a 32-bit ARM architecture? Using the command 'sudo pip install pyrealsense2' gives the error "Could not find a version that satisfies the requirement pyrealsense2 (from versions: ) No matching distribution found for pyrealsense2". Where can I get the package installed from?

[MartyG]

Did you use the Odroid-specific installation instructions to install Librealsense, please?

https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_odroid.md

[EYeo]

Yes i did follow that instructions, but however i cant seem to download the pyrealsense2 package

[MartyG]

It is a difficult question. I hope the link below will be of use to you.

https://github.com/IntelRealSense/librealsense/issues/2021#issuecomment-403722357

If you continue to have problems, the RealSense GitHub may be the best place to get help about an issue with 32-bit ARM.

[EYeo]

Hi, regarding the example you told me to follow from https://github.com/IntelRealSense/librealsense#ready-to-hack , is it possible to stream the camera while showing the distance? Cause I tried and there seems to be an error stating that the two programs cannot be run simultaneously.

[Alexandra_C_Intel]

Hi EYeo, Thank you for your interest in the Intel RealSense D435 camera. It is not possible to install the pyrealsense2 package onto a 32-bit ARM architecture, you will have to build it from source. https://github.com/IntelRealSense/librealsense/tree/master/wrappers/python#building-from-source Only x86 and AMD64 architectures are supported by pyrealsense2. https://pypi.org/project/pyrealsense2/#files Also, if you want to stream the camera while showing the distance you may need to combine SDK 2.0 with a person tracking module in the OpenCV software platform. https://github.com/IntelRealSense/librealsense/tree/master/wrappers/opencv/dnn Regards, Alexandra

[EYeo]

Yes I managed to run the dnn example which you have mentioned. But I would like to detect more object other than the ones provided in the dnn code. May I know how do I go about it? There isn't a need to recognize the object. I just have to detect that there is an object ahead. Because the objective of my project is to do obstacle avoidance.

[Alexandra_C_Intel]

Hi EYeo, Please feel free to change the code in order to suit your needs and share your results with the Community. Regards, Alexandra