EcoQuad

Automated AI-powered analysis tool for quantifying species coverage in photoquadrat images using deep learning.

What is it?

An AI tool designed to analyze photoquadrat images. It helps users measure species presence and percentage coverage. It uses a multi-model CV and automated image processing. Currently, EcoQuad is in development and supports analysis of two species: Rissoella and Chthamalus. Using advanced computer vision, it:

• Detects the photoquadrat frame automatically
• Identifies different species in the image
• Calculates coverage percentages per grid cell
• Generates reports and visualizations in seconds

Installation

Local Setup

1. Clone the repository

   git clone https://github.com/DolapoSalim/photoquadrats_analysis
   cd photoquadrats_analysis

2. Install dependencies

   pip install -r requirements.txt

3. Add models Create a models/ folder and place your YOLO models:

   models/
   ├── new_frame_detector.pt
   └── final_detector.pt
   

4. Run the app

   streamlit run app.py

File Structure

photoquadrats_analysis/
├── app.py                    # Main application
├── requirements.txt          # Python dependencies
├── models/
│   ├── new_frame_detector.pt
│   └── final_detector.pt
├── assets/
│   └── header.png
└── README.md

Model Requirements

You need two pre-trained YOLO models:

1. Frame Detection Model (new_frame_detector.pt)

   • Detects photoquadrat frame in images
   • Trained on quadrat photos

2. Species Segmentation Model (final_detector.pt)

   • Segments species in cropped frames
   • Identifies Chthalamus, Risoella, and other species

---

Made with ❤️ for ecological research.
Data used for production are intellectual properties of the ecology lab, Dept of Biology, University of Pisa