NatureSense

The Project

NatureSense is a plant and animal identification app utilizing machine learning for image recognition technology. The app has a home page that records and provides information about all of their observations. On the bottom navigation bar, there is a tab for both plant identification and animal identification. On these pages, the user can take a picture or select an image from their camera roll for identification. An identification page is then shown, displaying the identified animal or plant, an identification accuracy percentage, and information about the organism. On the community page, users can post information about their findings and connect with others. The app was made possible using Google Firebase and utilizing the Flutter SDK. Google Firebase is used to handle authentication and it allows for users to securely sign up and log in to their account. Firebase’s Realtime Database allows for the app to keep track of the plants and animals the user has identified. Flutter was used to create the app in the Dart programming language. TensorFlow Lite was used for machine learning and it integrates models trained on a large dataset of animals and plants with a large amount of training images. Once the image has been identified, a JSON file is used to get the information on the species found.

Mobile
Art
Education
Environment

Team Comments

I chose to make this project because...

I chose to make this project because I believe it is important to spread awareness about biodiversity and allow people to better understand nature. The app can help users identify potentially harmful plants and animals and provide valuable advice.

What I found difficult and how I worked it out

A difficulty was curating a diverse and comprehensive dataset of animals and plants. It was also difficult making the model accurately identify the plants and animals. To resolve this, a specific variety of plants and animals were chosen, and each had many training images.

Next time, I would...

Improvements would include writing scripts to pull animal and plant data for a more diverse database, running the training algorithms through more epochs in order to increase accuracy, improving community social platform and the information on the identification, and enhancing the user interface.

About the team

  • United States

Team members

  • Austin