Land Zoning Using Satellite Imagery and Machine Learning

EasyChair Preprint 15466

Abstract

Appropriate land use planning is one of the new frontiers of sustainable development, particularly in the context of fast developing cities. Most land use classification systems, especially those that rely on visual interpretation and rudimentary image processing, continue to face difficulties accommodating the sophisticated, extensive datasets produced by contemporary remote sensors. This study investigates the new techniques of land zoning utilizing space images in conjunction with machine learning algorithms, in this case, neural networks, for the automatic classification of various types of land and fitting them for the purposes of agriculture, urbanization and industrialisation.

In the project, a specific land cover type classification model has been created, trained, and tested with Sentinel-2 and location centered satellite data. By the use of high resolution images this approach allows classification of some vital land cover use categories i.e. barren land, vegetation, and water. In order to increase comprehensibility and guide further analysis, the classifications are superimposed on the respective images with every land type annotated and discussed. It shows how remote sensing data integrated with machine learning can solve land zoning issues efficiently, reliably and in a scalable manner helping to make economy and environment friendly land use decisions.

Keyphrases: Environmental Sustainability., Land zoning, machine learning, neural networks, remote sensing, satellite imagery, sustainable development, urban planning

@booklet{EasyChair:15466,
  author    = {Surya Vardhan Karella and R Reddy Neeraj and M Aruna},
  title     = {Land Zoning Using Satellite Imagery and Machine Learning},
  howpublished = {EasyChair Preprint 15466},
  year      = {EasyChair, 2024}}