Thermal Camera and Distance Integration Project

Index

• Overview
• Getting Started
  • Prerequisites
  • Dependencies
  • Installation
• Usage
  • Selecting ROI
  • Running the Script
  • Saving Data
• System Overview
  • Circuit Schematics
  • Hardware Development
  • Software Overview
  • Communications Overview
• Limitations and Future Improvements
• Demo
• Authors
• References

---

Overview

This project combines a thermal camera and an ESP32 to capture and analyze thermal data and distance measurements. It was developed as part of the YAKISUGI Torch Project, a system designed to optimize and standardize the traditional Japanese wood-burning technique for enhanced durability, pest resistance, and waterproofing.

About the YAKISUGI Method: The YAKISUGI method is a traditional Japanese wood preservation technique that involves charring the surface of the wood with fire. This process results in increased durability, pest resistance, and waterproofing.

Objective and Motivation:

• Optimize the wood-burning process by providing visual feedback, measurements, and real-time data.
• Preserve and modernize this ancestral technique for artisans and robotic systems.
• Enable interaction between digital devices and artisans, informing future designs for robotic systems.

The thermal camera streams real-time data, allowing users to select a Region of Interest (ROI), while the ESP32 provides proximity data. The script integrates both sources of information, displays them on an interactive interface, and saves them to a CSV file.

Main Features:

• Real-time thermal data processing.
• Integration with an ESP32 for distance measurement.
• Interactive user interface for ROI selection.
• Data logging of temperature and distance values.
• Preservation of the YAKISUGI technique through standardized and optimized burning processes.

---

Getting Started

Prerequisites

Ensure the following are installed:

• Ubuntu 20.04 LTS or later.
• Python 3.7 or newer.
• Compatible Thermal Camera (In our case PureThermal 2)

Dependencies

Install the required Python libraries:

pip3 install numpy opencv-python requests

Additional system libraries might be needed depending on your thermal camera:

sudo apt-get install libuvc-dev

Dependencies

Clone this repository to your local machine, and follow the instructions to install the libraries to use the camera.:

git clone https://github.com/Clarrainl/thermal-camera.git
cd thermal-camera

Installation

1. Clone this repository:

git clone https://github.com/MRAC-IAAC/YAKISUGI.TORCH
cd YAKISUGI.TORCH

2. Verify your ESP32 is set up and running with its provided firmware.

   • The ESP32 should serve distance data on a network-accessible IP (default: http://192.168.4.1).

3. For additional guidance on using the thermal camera, refer to the repository developed for this purpose: Thermal Camera GitHub Repository.

---

Usage

Selecting ROI

1. Run the script:

   python3 yakisugi_python.py

2. The live feed from the thermal camera will appear.

3. Press s to select the Region of Interest (ROI):

   • Use your mouse to drag and select a rectangular area.
   • This area will be analyzed for temperature data.

Running the Script

• Once the ROI is selected, the script will:

  • Continuously process thermal data within the ROI.
  • Query the ESP32 for distance measurements.
  • Display the maximum and minimum temperatures, as well as the distance, in real-time.

• Press q at any time to exit the program.

Saving Data

• The script logs temperature and distance data at regular intervals (default: every 5 seconds).
• After exiting, you will be prompted to save the data to a CSV file:
  • Example output file: temperature_distance_data.csv.

---

System Overview

The system integrates various hardware and software components to achieve real-time feedback and optimization of the wood-burning process:

Hardware Components:

• Thermal Camera: FLIR Lepton (PureThermal 2 Cam) for capturing temperature data.
• ESP32: Microcontroller for distance measurement and Wi-Fi communication.
• Ultrasonic Sensor: HC-SR04 for proximity sensing.
• OLED Screen: 0.96-inch Waveshare for real-time display of temperature, time, and distance.
• Battery and Voltage Regulator: Power supply for the ESP32 and peripherals.

Software Features:

• Thermal Data Processing: Captures and processes temperature data, converting it into Celsius and generating color-mapped thermal images.
• Distance Measurement: Queries the ESP32 via HTTP to retrieve real-time distance data.
• Real-Time Visualization: Displays temperature and distance on an interactive interface, allowing artisans to optimize the burning process.
• Data Storage: Logs time, temperature, and distance in a CSV file for post-process analysis.

Circuit Schematics

This system integrates multiple hardware components:

1. Ultrasonic Sensor (HC-SR04): Measures distance to ensure proper torch positioning.
2. ESP32 Microcontroller: Manages data from the ultrasonic sensor and streams it to Python via HTTP.
3. OLED Screen: Displays real-time data to the user.
4. PureThermal Camera: Captures and streams thermal data.

---

Hardware Development

Working Principles:

• Thermal Capture: Measures surface temperature of the wood using a thermal camera.
• Distance Measurement: Ensures consistent torch positioning through an ultrasonic sensor.
• Data Processing: ESP32 processes time, temperature, and distance data and displays it in real-time on the OLED screen.

Software Overview

Libraries Used:

• OpenCV & NumPy: For image processing and numerical computation.
• uvctypes: Communication with the thermal camera.
• Queue: Manages real-time thermal data buffering.
• CSV: Logs processed data for later analysis.
• Requests: Queries the ESP32 for distance measurements via HTTP.

Communications Overview

Protocols Used:

• USB Video Class (UVC): Streams thermal frames from the PureThermal camera.
• HTTP: ESP32 hosts a server, and Python retrieves distance data in JSON format.
• Trigger-Echo: Ultrasonic sensor sends pulses and calculates distance using timing.

---

Limitations and Future Improvements

Limitations

• The resolution of the thermal camera limits precision for detailed temperature mapping.
• Ultrasonic sensors are sensitive to surface inconsistencies, affecting accuracy.
• The OLED screen size constrains the amount of real-time data displayed.

Future Improvements

• Upgrade to a higher-resolution thermal camera for detailed mapping.
• Replace the ultrasonic sensor with LiDAR for enhanced precision.
• Integrate data with mobile or desktop applications for better visualization.
• Add ambient condition sensors for humidity or other environmental factors.
• Implement audio alerts to enhance safety during burning.

---

Demo

To demonstrate the script:

1. Connect your thermal camera and ESP32.
2. Run the script as described in Usage.
3. Observe the real-time feed and the logged data in the CSV file.

The results include a table with:

Time (HH:MM:SS)	Min Temp (C)	Max Temp (C)	Distance (cm)
00:00:05	20.5	37.8	15

---

Authors

• Charlie Larraín - Developer
• Javi Albo - Developer
• Mau Weber - Developer

This project was developed as part of the Master in Robotics and Advanced Construction (MRAC), Term II, 2024/25.

---

References

• FLIR Camera Documentation
• ESP32 Documentation
• Thermal Camera GitHub Repository

---