We use Computer Vision technology to identify urban heat islands. This completely automates the process
In the status quo, scientists and researchers around the world studying urban heat islands and their effects are identifying and indexing urban heat islands manually, costing them countless of lost hours and leaving room for error. CHILL AI eliminates the need for this.
We're implementing the YOLOv8 Computer Vision model, which is known for its fast detection speeds and real-time capabilities. This allows any new information that is inputted to CHILL AI to be processed within seconds and allow a live overview of the state of urban heat islands to be seen.
Currently, there are no standardized, widely-used systems that can quantitatively measure the magnitude of urban heat islands. However, through CHILL AI, all insights are classified under the same weights and biases developed during the training process derived the training/validation/testing dataset. This means that classfications made for the severity and priority of action for urban heat islands are standardized, giving governments and organizations an objective overview of them.
The model we're implementing, YOLOv8, is one of the most widely used and studied Computer Vision pre-training models. Thousands of different use cases have been applied for it, ranging from simple applications for drawing bounding boxes around different types of bottles to highly complex applications such as scanning individuals for hidden firearms and detecting their specific make. Its accuracy and results in completing these tasks have proven to reach accuracies of 95% and above as well, consistently.
Geographic Information Systems data will be sourced from government databases and organizations such as the US Geological Survey (USGS) and Esri.
GIS platforms provide detailed spatial data on land use, infrastructure, and demographics.
Satellite data will be obtained from NASA’s Earth Observing System Data and Information System (EOSDIS) and the MOderate Resolution Imaging Spectroradiometer (MODIS) aboard the Terra and Aqua satellites.
These sources offer high-resolution thermal imagery essential for monitoring land surface temperatures.
Data from mobile weather applications and networks of IoT sensors will be used.
Companies like Weather Underground and community-based networks like PurpleAir provide real-time weather and air quality data from many sensors distributed across the city/urban areas.
