Shriya Gautam

Using Computer Vision to analyze Satellite Images for Ecological Characteristics

Started May 17, 2022

Abstract or project description

My paper will utilize computer vision technology, specifically convolutional neural networks, to analyze satellite images for visual characteristics. My convolutional neural network(CNN) will identify visual characteristics within the satellite images such that it can determine which species of plants and/or animals are present in the photo. For example, given a satellite image of the Amazon rainforest, the network would be able to identify the species of trees present in that specific area. After identifying these features, the CNN will feed these features to a logistic regression model that would use the features to determine which ecosystems across the world would have a similar makeup of plants. Current research in this field has allowed farmers and scientists to identify plant health and project future growth using satellite images of areas like forests and fields. This paper would build off of existing research about analyzing satellite imagery to more efficiently make conclusions about certain areas using photos instead of in-person investigations.