Gen6mini Vision System & Simulation

Overview

This repository contains the ROS 2 (Jazzy) packages for the Gen6s robot platform. It focuses on differential drive simulation (Gazebo) and visual perception, specifically Optical Character Recognition (OCR).

Packages

Package	Description
test_diff_robot	Contains the URDF description and Gazebo simulation launch files for the Gen6s differential drive robot.
gen6s_character_detection	Performs OCR using Tesseract and custom template matching to read text (e.g., "STOP") from image streams.
gen6s_color	A testing utility that publishes static images (like a Stop Sign) to a ROS topic to mock a camera feed.
gen6s_bnw	Visualizes the image processing pipeline by converting color inputs to binary/black-and-white.
gen6s_description	Basic robot description package.

Prerequisites

• OS: Windows (WSL2) or Ubuntu 24.04 (Noble Numbat)
• ROS 2: Jazzy Jalisco
• Dependencies:
  • tesseract-ocr
  • libtesseract-dev

Installation

1. Install System Dependencies

   sudo apt update
   sudo apt install tesseract-ocr libtesseract-dev

2. Build the Workspace

   # From your workspace root (e.g., ~/ros2_ws)
   rosdep install --from-paths src -y --ignore-src
   colcon build --symlink-install
   source install/setup.bash # or .ps1 for PowerShell

Usage

1. Run the Robot Simulation

Launch the robot in Gazebo and RViz to see the physical model.

ros2 launch test_diff_robot test_diff_robot.launch.py

Note: If the robot model is missing, check the xacro properties in urdf/robots/test_diff_robot.urdf.xacro.

2. Run the Vision/OCR Pipeline (Test Mode)

Since the simulation camera is not yet linked to the OCR node, use the mock publisher to test text recognition.

Terminal 1: Publish the Test Image

ros2 launch gen6s_color gen6s_color_launch.py

Publishes a static image to /imu/data.

Terminal 2: Run OCR Node

ros2 run gen6s_character_detection char_recognition_node

Subscribes to /imu/data and interprets the text.

3. Run the Image Process Visualization

To see how the computer "sees" the image (Binary Thresholding):

ros2 launch gen6s_bnw gen6s_bnw_launch.py

Architecture Notes

• Topic Misnomer: The vision nodes currently communicate over the topic /imu/data instead of standard camera topics effectively treating it as a raw image stream.
• OCR Method: The system uses a hybrid approach of Tesseract (LSTM) for full phrases and Template Matching (OpenCV contours) for individual characters.