Animal Detector

• 

  Problem / Solution

• 

  Training the AI Model

• 

  Detect Deer

• 

  Log of animal encounters and sounds emitted

• 

  Detect Opposum

• 

  Detect Aramadillo

• 

  Camera feature and GPT-Engineer IDE

Problem: There is a need to reduce the number of vehicle collisions with wildlife to prevent injuries and preserve endangered species.

Solution: Ultrasonic AI animal repeller device.

• Computer vision to identify animals
  • Ultrasonic audio to scare detected animals
  • Blinking light to alert drivers
  • Integrates into roadside infrastructure

For this prototype, image datasets of animals from Kaggle were used to classify 8 groups of animals commonly found near the road. These images were uploaded to Teachable Machine to train an AI model to identify animals. The TensorFlow trained model was integrated into a GPT-Engineer built react app, allowing a user to upload or take photos with a device for the AI to recognize, and emit an ultrasonic sound. A log records the animals detected and the ultrasonic audio that was emitted.

Built With

• gpt-engineer
• javascript
• react
• tensorflow