Classification of Flying Objects with Computer Vision and Deep Learning Integrated System

The study has been tailored that the address the problem in aerial objects domain. To tackle down a solution to provide state of art method on the domain of drone and aerial object classification that helps vast majority of countries’ defense systems to provide a vital defense in critical circumstances. Deep Learning (DL) and Computer Vision (CV) integrated system has been provided a state of art solution to tackle down the problem. Convolutional Neural Network (CNN), DenseNet, MobileNetV2, ResNet50, VGG16, VGG19 models have been utilized to demonstrate a classification of the images that the image set consists of training, testing and validation splits of drone|non-drone objects. Outcomes have been stated that in many performance metrics DenseNet has been outperformed other models; in contrast ResNet50 and VGG16 has been provided bad outcomes in several performance metrics that the authors have been suggested to use AlexNet in demonstration of classification problem in Unmanned Aerial Crafts domain.