How to Use YOLO v8 with ZED in Python

Introduction #

This sample shows how to detect custom objects using the official Pytorch implementation of YOLOv8 from a ZED camera and ingest them into the ZED SDK to extract 3D informations and tracking for each objects.

Installation #

ZED Yolo depends on the following libraries:

  • ZED SDK and [Python API]
  • Pytorch / YOLOv8 package
  • OpenCV
  • CUDA
  • [Python 3]


Install the ZED SDK and Python API.

Setting up #

  • Install yolov8 using pip
pip install ultralytics

Run the program #

NOTE: The ZED v1 is not compatible with this module

python --weights # [--img_size 512 --conf_thres 0.1 --svo path/to/file.svo]

Features #

  • The camera point cloud is displayed in a 3D OpenGL view
  • 3D bounding boxes around detected objects are drawn
  • Objects classes and confidences can be changed

Training your own model #

This sample can use any model trained with YOLOv8, including custom trained one. For a getting started on how to trained a model on a custom dataset with YOLOv5, see here

Support #

If you need assistance go to our Community site at