depth-vision

ZED-2 AI Stereo Camera

By
0

In a few words... this is a great camera and ecosystem (you need) if you start the prototyping phase of the product in the early stage.

Stereolabs delivers a solid CUDA enabled SDK with plenty of examples, wrappers, and integrations. It's also well optimized for the Jetson family.

Unfortunately, I have experienced some issues with the compilation of OpenPose or OpenCV wrappers & examples. In most of the cases, it's due to bugs in 3rd party dependencies (you might need some workarounds and hacks, too).

In my case, the camera is used with Jetson Nano 4GB, Jetson TX-1, and Jetson Xavier AGX. It works perfectly well on TX-1 and Xavier, while Jetson Nano can experience performance problems in higher resolutions (due to limited CUDA throughput and slow 4GB memory).

Point Cloud from ZED-2
ZED-2 and OpenPose integration

Comments

Related posts

stereo-vision

RealSense D455 Stereo Camera from Intel

The new arrival in my home-lab. The Intel RealSense D455 stereo camera - just freshly released. It's made well and works out of the box. Next week will be testing the accuracy of the depth map and also the quality…

Chihuahua

Art 3D/VR challenge – cameras and sensors mount

I have designed a very first prototype of camera mount which could be used in the VR project. It is very rough construction, but the idea is that: Below some photos of my design. There is still missing camera, servo…

3D

Art 3D/VR challenge – waiting for camera

Introduction I was researching the right cameras for very long time. The market is full of products on every level: I have found a plenty of producers and I did research on following: This is my very abstract drawing of…