Eco Miami Watch

An Effort to Environmental Conservation. Development of User-Friendly Web App and AI in Remote Sensing for Water Pollution Monitoring and Deforestation Monitoring in Miami-Dade.

Comment

Inspiration

Miami faces multiple environmental challenges due to rapid urbanization, water pollution, and deforestation which leads to Sea-level rise and coastal erosion, Pollution and urban runoff affecting water quality, Increased frequency of extreme weather events, Habitat loss impacting local biodiversity.

What it does

  • Streamlined Monitoring of Miami’s Natural Environment
  • Water Pollution Monitoring 2. Deforestation Monitoring

How we built it

We built it with Python, PyTorch, Sam2 pretrained model, Lovable, Colab, Leafmap, Google Engine, Numpy, Copernicus Data Space APIs, geemap, Hugging Face Random Forest Model, SamGeo

Challenges we ran into

  1. Unavailability of Proper Spectral Band of Satellite Imagery
  2. Difficulty in Finding the Critical Deforestation Area in Miami-Dade
  3. Proper API to the Satellite Imagery
  4. Difficulty in Collection of Dataset
  5. Finetuning Random Forest with the available data

Accomplishments that we're proud of

  1. We are able to develop UI for the Monitoring System to be used in Miami-Dade
  2. Finetuned ML model for detecting water pollution
  3. Applied SamGeo for the segmentation of critical area in Miami-dade regarding Forest Deforestation
  4. Able to determine if there is deforestation with pixel analysis in Satellite Imagery

What we learned

Samgeo, Lovable, Extraction of data from the geemap, usage of HuggingFace, Google GIS system,

What's next for Eco Miami Watch

Enhance accuracy by counting the exact number of trees. Improve segmentation by fine-tuning SAMGeo on a larger dataset. Facilitate extraction of satellite imagery for specific dates. Develop real-time data processing capabilities. Strong Water Quality Monitoring

Built With

Share this project:

Updates