INTELLIGENT LAND COVER CLASSIFICATION AND ROUTE PLANNINGUSING HYPERSPECTRAL-LIDAR FUSION AND OPTIMIZER

Abstract

The proper location of transmission lines is critical for optimising energy distribution while minimising environmental damage. This research seeks to improve the identification of potential places for transmission line placement by using sophisticated classification algorithms, notably Convolutional Neural Networks (CNNs) based on hyperspectral and LiDAR data.

We used a comprehensive strategy that included several remote sensing technologies to properly categorize land cover types and provide a binarized picture showing potential installation locations. The system featured a rigorous CNN-based categorisation procedure, followed by optimum route selection for transmission line routing, which ensured environmental compliance.

Our findings show good classification accuracy across a variety of land cover categories, with particularly strong performance in "Trees," "Road," and "Water," with accuracy rates of 98%, 95%, and 97%. The misclassification rates were also analyzed, giving information for future model modification. The resulting visualizations clearly depict the ideal transmission line pathways, demonstrating the model's trustworthiness.

In conclusion, this research effectively creates a robust framework for transmission line planning that uses CNNs and remote sensing data to deliver actionable insights, therefore contributing to long-term infrastructure development.