Assigning accuracy for Cars Detection from High-Resolution Satellite Images Using Different Machine Learning Models

Abstract

High-resolution satellite imagery provides a wealth of detailed visual information that can be leveraged for various machine learning applications. In this study, we present a deep learning-based approach for car classification using high-resolution satellite images. Utilizing the powerful capabilities of Tensor Flow layers, we design and implement a convolutional neural network (CNN) to accurately identify and classify different types of cars from satellite imagery. The process involves the collection of a diverse dataset of satellite images containing vehicles, followed by rigorous data pre-processing and augmentation to enhance model robustness. The CNN architecture is optimized through hyper parameter tuning and trained on a labeled dataset, achieving high accuracy in classifying vehicles into predefined categories such as sedans, SUVs, and trucks. Our results demonstrate the effectiveness of using deep learning models with TensorFlow layers for car classification tasks, highlighting the potential for broader applications in urban planning, traffic management, and automated vehicle detection from satellite imagery.