Satellite Image Processing for Land use and Land Cover Mapping

  • Unique Paper ID: 177715
  • Volume: 11
  • Issue: 12
  • PageNo: 1197-1202
  • Abstract:
  • In this paper, we present GeoClass, a Python-based satellite image classification system that leverages deep learning to automate land use analysis and scoring. Traditional land evaluation processes are limited by manual data collection and interpretation. GeoClass addresses these limitations by employing a pre-trained EfficientNetB0 model, fine-tuned for satellite imagery, to classify land parcels into ten categories such as forests, lakes, highways, and industrial zones. The system processes multi-spectral images through normalization, resizing, and augmentation to enhance training performance and generalization. Classification is performed with softmax-based probability outputs, enabling confidence scoring and top-class prediction visualization. A unique feature of GeoClass is its land scoring module, which evaluates land value based on proximity to features like water bodies, highways, and industrial zones using weighted distance maps. Designed for scalability and accuracy, GeoClass enables batch classification in grid format and provides heat map visualizations of land values, supporting applications in real estate, urban planning, and environmental monitoring. By automating classification and value assessment, GeoClass offers a powerful, efficient solution for modern land resource analysis.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 12
  • PageNo: 1197-1202

Satellite Image Processing for Land use and Land Cover Mapping

Related Articles