BirdCamAI
BirdCamAI
Embedded Files
Developed a real-time bird detection and identification system capable of accurately detecting and classifying approximately 900 bird species with an average latency of 300ms. Upon detection, the system automatically sends an email alert.
This project utilizes a Raspberry Pi camera and an ML model running in Pi Zero 2W, overcoming challenges such as varying lighting conditions and bird occlusion.
BirdCamAI Prototype  Sample Email
Currently...
Developing a bird detection model using Google Open Images Dataset V7 and YOLO11 architecture on AWS gpu based EC2 Instance (g5.xlarge) and s3.
Long-term goals include:
IoT Integration: Develop an IoT product by storing captured images in the cloud, enabling remote access and monitoring.
Model Enhancement: Expand the model to detect a wider range of bird species with improved accuracy and reduced latency.
Data-Driven Applications: Utilize the collected data to develop valuable insights and data sources for ecological research and conservation efforts.
Stay tuned for more updates !!!
Page updated
Google Sites
Report abuse