PetAnalytics

This repository contains multiple hardware and software projects, all part of the PetAnalytics project. Each subproject serves a unique purpose, ranging from IoT devices to video annotation software.

1. Sparkfun Projects

This project focuses on developing an IoT device to collect data on animal behavior using an ESP32 SparkFun Thing Plus and SparkFun 9DoF IMU Breakout - ICM-20948 sensors.

• Features:
  • Real-time motion data collection from the IMUs.
  • Data logging to an SD card.
  • Configurable IMU sample rates for various tracking scenarios.
  • Potential for integration with a local database on the SD card.
  • Expandable system to support multiple IMUs simultaneously.

Subprojects

• One IMU (sdcard/one): Project using a single IMU sensor to capture motion data and store it on an SD card.
• Three IMUs (sdcard/three): Version with three IMU sensors for more comprehensive data collection and SD card storage.
• Local Database (local_database): Implementation of a local database on the SD card to store and retrieve collected data. (Needs review)

2. M5Stack Projects

Projects utilizing various M5Stack hardware for data capture and monitoring. M5 devices are practical and modular, allowing easy integration with a variety of sensors.

• Features:
  • Modular design allowing rapid prototyping.
  • Integration with Firebase for real-time data storage
  • Support for SD card storage, allowing local data logging
  • Compact and portable, ideal for field data collection.

Subprojects

• Capsule: Project using the M5Stack Capsule, a compact device for data collection and monitoring.
• StickCPlus: Project using the M5StickC Plus, a small but powerful IoT development kit.
• Multiple IMU (multiple_imu): Tests with external IMUs using the M5 IMU units for motion data collection from various sensors. (To-Do)

3. Video Annotation Software

This software is designed for video annotation and labeling, specifically to label animal behavior. It provides a graphical interface for video manipulation, including playback at different speeds and frame-by-frame labeling.

• Technologies used: Python, Tkinter.
• Features:
  • Frame-by-frame video labeling.
  • Flexible playback speeds (0.25x, 1x, 2x).
  • Keyboard shortcuts for applying labels.
  • Export of labeled data to CSV format.
  • Synchronization with sensor data (IMU readings) (To-Do).

Acknowledgements

Special thanks to the team members, Professor Rafael de Pinho André, and FGV EMAp for their contributions, support, and resources provided to this project.

Contact

For any inquiries or feedback, please contact us at: [email protected]