Abstract
This study investigates the potential of using deep learning for real-time image analysis in assessing sustainable infrastructure and urban development. Convolutional Neural Networks (CNNs) are implemented to evaluate live-captured building images, enabling automated classification and data extraction for decision-making. The proposed approach overcomes the limitations of existing methods by facilitating real-time analysis and large-scale data processing. A dataset exceeding 12,000 images rigorously evaluates the CNN model's performance. This research establishes a framework for leveraging deep learning for real-time assessment of sustainable infrastructure, paving the way for improved data-driven urban planning and development decision-making. The study confirms that the Inception Net V3-based feature extraction technique accurately classifies images based on CO2 emission levels. This classification task is best performed using the Neural Network model. Advanced feature extraction techniques are essential for improved environmental image classification.