ROS2 Depth Anything V3 TensorRT
ROS2 TensorRT node for monocular metric depth
Description
A robot with only an ordinary camera lacks depth and point clouds. This project provides a ROS2 TensorRT node for monocular metric depth with Depth Anything V3.
It estimates real-scale depth and point clouds from single images in real time on NVIDIA hardware.
Mono depth:Metric scale.
Point clouds:Real time.
TensorRT:Fast inference.
ROS2:Drop-in.
It estimates real-scale depth and point clouds from single images in real time on NVIDIA hardware.
Features
Mono depth:Metric scale.
Point clouds:Real time.
TensorRT:Fast inference.
ROS2:Drop-in.
