OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Darknet Ros

Darknet Ros is a C++ ROS package that runs YOLOv3 object detection on camera images using CPU or GPU inference. Its main node consumes ROS camera measurements and publishes detected-object counts, bounding boxes and an annotated detection image, with support for pretrained VOC or COCO classes and custom detection networks.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Darknet Ros — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a ROS application needs YOLO-based detections from camera images exposed through ROS topics or an image action.

Who it is for

  • ROS developers integrating object detection into robot camera pipelines.
  • Researchers or developers using pretrained VOC or COCO classes in camera imagery.
  • Teams adapting a custom YOLO network for ROS image processing.

Documented uses

  • Process a ROS camera image stream with YOLOv3.
  • Make detected-object counts, pixel-coordinate bounding boxes and annotated images available to ROS components.
  • Select a pretrained model or configure custom network files for detection.
Before you choose

Requirements & limits

The repository is public and specifies the BSD-3-Clause license.

  • The documented build depends on ROS, OpenCV and Boost and uses a catkin workspace.
  • The README reports testing with ROS Noetic on Ubuntu 20.04 and mentions branches for ROS Melodic, ROS Foxy and ROS2.
  • CPU inference is supported; GPU compilation requires an Nvidia GPU and CUDA.
  • The node's documented outputs include object counts, bounding boxes and a detection image.
  • The project describes itself as research code that may change.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Build Darknet Ros in a catkin workspace

Prepare and compile the package in a ROS catkin workspace.

Before you start

  • ROS installed
  • A catkin workspace
  • OpenCV and Boost
  • An SSH key configured for cloning the repository

Steps

  1. From the workspace source directory, clone the repository recursively using its documented SSH URL.

  2. Return to the workspace root.

  3. Compile in Release mode with catkin_make or Catkin Command Line Tools.

Expected outcome

The package is compiled in the catkin workspace.

Guide limitations

  • The README reports testing with ROS Noetic and Ubuntu 20.04; it also lists branches for ROS Melodic, ROS Foxy and ROS2.
  • GPU compilation requires an Nvidia GPU and CUDA; the README says the build switches to CPU when CUDA is not installed.
  • The README describes the project as research code that may change.
Documentation 1 ↗
Source-based practical guide

Configure a robot camera and detection network

Adapt package configuration and launch files for a robot camera and selected YOLO model.

Before you start

  • A built Darknet Ros package in a ROS workspace
  • A ROS camera image topic
  • A network configuration file and corresponding weights in the package's documented directories

Steps

  1. Duplicate and adapt the necessary parameter files in the package's config directory and the launch file in its launch directory.

  2. Set the ROS topic and related publisher, subscriber or action names in ros.yaml as needed.

  3. Configure the network configuration and weights filenames in the detection parameters.

Expected outcome

The package is configured to subscribe to camera images and select a detection network; the node is documented to publish object counts, bounding boxes and a detection image.

Guide limitations

  • The README describes adapting configuration but does not provide a complete robot-specific launch configuration.
  • Custom detection requires the user's own network configuration and weights.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
leggedrobotics/darknet_ros · Official homepage title · retained in its source language
How we review sourcesSuggest a correction