Automated Urban Infrastructure Classification Using Aerial Imagery

Automated classification of urban infrastructure from aerial imagery is crucial for urban planning, environmental monitoring, and disaster response. However, conventional methods often face challenges with accuracy and scalability. This study presents an innovative approach that combines the EfficientNetB3 Convolutional Neural Network (CNN) with Long Short-Term Memory (LSTM) networks to improve classification accuracy for urban infrastructure elements. We trained our model over 40 epochs with a batch size of 51, employing an 80:20 split for training and testing. Our method achieved a classification accuracy of 96% on a comprehensive dataset, surpassing existing approaches. Notably, the model attained perfect precision and recall scores of 1.00, demonstrating exceptional performance in classifying diverse urban infrastructure components. The findings suggest that this approach has significant potential in urban planning, disaster response, and environmental monitoring. Future research will focus on testing the model's scalability with larger datasets and exploring its robustness under different environmental conditions.