A ROS 2 Wrapper for Florence-2: Multi-Mode Local Vision-Language Inference for Robotic Systems

TL;DR

This paper introduces a ROS 2 wrapper for Florence-2, enabling multi-mode vision-language inference in robotic systems with local execution on consumer hardware.

Contribution

It provides a flexible, multi-mode ROS 2 interface for Florence-2, facilitating practical integration of advanced vision-language models into robotic software stacks.

Findings

Feasible local deployment on consumer-grade GPUs.

Supports three interaction modes: topic, service, action.

Repository available at https://github.com/JEDominguezVidal/florence2_ros2_wrapper.

Abstract

Foundation vision-language models are becoming increasingly relevant to robotics because they can provide richer semantic perception than narrow task-specific pipelines. However, their practical adoption in robot software stacks still depends on reproducible middleware integrations rather than on model quality alone. Florence-2 is especially attractive in this regard because it unifies captioning, optical character recognition, open-vocabulary detection, grounding and related vision-language tasks within a comparatively manageable model size. This article presents a ROS 2 wrapper for Florence-2 that exposes the model through three complementary interaction modes: continuous topic-driven processing, synchronous service calls and asynchronous actions. The wrapper is designed for local execution and supports both native installation and Docker container deployment. It also combines generic JSON outputs with standard ROS 2 message bindings for detection-oriented tasks. A functional validation is reported together with a throughput study on several GPUs, showing that local deployment is feasible with consumer grade hardware. The repository is publicly available here: https://github.com/JEDominguezVidal/florence2_ros2_wrapper