Forest Scan: An Image Analytics for Tree Enumeration for the Diversion of Forest Land

Forest Scan is an innovative Image Analytics model designed to address the critical need for accurate tree enumeration in regions designated for forest land diversion for the infrastructure development. Forest Scan uses advanced technology, including satellite imagery, aerial photographs, and drone inputs, to automate the tree enumeration process so that to make the ecological survey speedy. The model focuses on vital aspects like accurate counting and categorization of trees, rigorous validation procedures, scalability, efficiency, integration, and visualization. Forest Scan boasts the unique capability of employing two distinct models for tree counting and species classification. Using YOLO (You Only Look Once) model it accurately counts trees with accuracy of 85%, while the Convolutional Neural Network (CNN) model excels in classifying trees into six different species, including palm, spruces, oaks, pine, banayan and sandalwood with accuracy of 70%.