Deep Neural Network Based Vehicle Detection and Classification of Aerial Images

The detection of the objects in ariel image has a significant impact on field parking space management, traffic management activities and surveillance systems. Traditional vehicle algorithms have some limitations as these are not working with complex background small size object bigger scenes. It is observed that researchers facing numerous problems classification, i.e., complicated background, vehicle’s modest size, other similar visual appearances correctly addressed. A robust algorithm for classification been proposed to overcome limitation existing techniques this research work. We propose an based Convolutional Neural Network (CNN) detect classify it into light heavy vehicles. performance approach was evaluated using variety benchmark datasets, including VEDAI, VIVID, UC Merced Land Use, Self database. To validate results, various parameters such accuracy, precision, recall, error, F1-Score were calculated. results suggest technique higher rate, which approximately 92.06% VEDAI dataset, 95.73% VIVID 90.17% 96.16% dataset.