Welcome to the Git repository for the Sensor CDT's 2024 Team Challenge: Automated Vegetation Monitoring Using Animal-Mounted Sensors.
You can find out about each component used in DAISY via the links below.
- Seeed Studio XIAO ESP32S3 Sense + Expansion Board
- OmniVision OV5640 CMOS Image Sensor
- SanDisk 32GB High-Endurance microSD Card
- JLCPCB Flexible Printed Circuit Board
- RS PRO 18650 Lithium-Ion Rechargeable Batteries
- Kvikk Durable Collar for Cattle
To compile our code, please include the following libraries through the Arduino IDE's Library Manager. Furthermore, the board should be set to XIAO_ESP32S3, the correct COM port must be selected and Espressif's ESP32 board package should be installed under File->Preferences.
- SD
- OV5640 Auto Focus for ESP32 Camera
- SparkFun LIS2DH12 Arduino Library
- SparkFun LIS3DH Arduino Library
- SparkFun u-blox GNSS v3
All dataset-related scripts and data can be found in the Dataset directory.
The raw JPEG files of the image fragments used in the dataset can be downloaded here.
The ML model implementation, experiment scripts, notebooks and HPC configurations can be found in the Model directory.
You can download the weights (model.pth file) for our best ML models using the links below.
- Best individual models: iNaturalist, DenseNet121
- Best ensemble model
To find out more about this project, please visit our website (daisysensing.com) or Instagram profile (instagram.com/daisysensing).