Neural Network Model for Monitoring of Landfills Using Remote Sensing Data

The stockpiling and retention of waste is an actual problem for Ukraine today. Count of illegal dumps increases every day because many landfills are full. Special services of waste monitoring provide only partial information about the current state of waste dumps all over the country. They can’t analyse landfills’ size information and their changes over time. Authors create an efficient algorithm for waste monitoring that uses historical and actual satellite data. It provides information about the current area of landfills and makes it possible to analyse changes in landfills over time. The authors found success resolving the main problem of separating landfills from a quarry and artificial objects. This solution is based on mixed using pixel-based and object-based classification.