Conclusion

This project successfully developed an AI-based system for detecting and counting rabbits in a farm environment using advanced object detection techniques. It was a multifaceted endeavor, requiring meticulous planning, experimentation, and iterative problem-solving to achieve the desired results.

Summary of Work

The project spanned several phases:

• Data Preparation: Extracting frames from video footage, annotating them using LabelImg, and splitting the dataset into training, validation, and test sets.

• Model Training: Leveraging TensorFlow and TFLite to develop an optimized object detection model capable of real-time rabbit detection.

• Deployment: Testing the model on video data and building cross-platform applications for practical use, including macOS and Windows.

Key Challenges and Solutions

Major challenges included computational limitations, handling large datasets, and building platform-specific apps. These were addressed through platform selection (Google Colab Pro), leveraging university storage solutions, and cross-platform testing using Azure VMs and high-performance hardware.

Current Achievements

• The project is now deployed in a real-world setting, with the AI system actively used for monitoring rabbits in a farm environment.

• A fully functional app for both macOS and Windows has been delivered, facilitating easy use by non-technical stakeholders.

Future Directions

• Model Enhancements: Incorporate additional datasets and refine hyperparameters to further improve detection accuracy and efficiency.

• Documentation and Open Source: Finalize and upload the entire codebase with detailed documentation, demo videos, and GitHub integration for wider accessibility.

• Scalability: Explore cloud-based solutions to handle larger datasets and environments.

This project has been a remarkable learning journey, blending theoretical knowledge with practical challenges. By overcoming these challenges, the system is now a step closer to advancing the integration of AI into agricultural practices, paving the way for future innovations in automated animal monitoring.