Hello. I am working on an autonomous robot that has to define the distance between itself and nearby objects. Also, it has to recognize some shapes.
Would you rather recommend to use a dedicated depth camera, like Stereolabs ZED, Intel RealSense Depth camera, or use two cameras to make 3D map?
The main question is whether the second option (two cameras) is possible and reliable.
Thank you in advance.
Hello. I do connected more than 2 cameras to Jetson Nano. If you want to connect 2 webcams - make sure that they got their own voltage regulator, which does not burdens Jetson’s power source. Got my Jetson burned with HDMI 5" screen connected to Jetson’s USB for power.
Thank you. Power is not the issue. I am trying to figure out whether i can use two separate cameras in order to create a real-time 3D map of surroundings.
Here is a video of what I want to achieve.
Hello. Why not? I do image threshholding in OpenCV from 5 to 250 and almost got what you needed with one camera (image threshholding and painting detected countours on output image, try it). So, if you got two cameras, ( make sure to use NVMM memory to accelerate process, why else you need Jetson Nano :-) ) you can make any dot-detection routine and use dots to make 3D map of objects.