Multiple Objects and Road Detection in Unmanned Aerial Vehicle

Unmanned Aerial Vehicles have greater potential to widely used in military and civil applications. Additionally equipped with the cameras can also be used in agriculture and surveillance. Aerial imagery has its own unique challenges that differ from the training set of modern-day object detectors, since it is made of images of larger areas compared to the regular datasets and the objects are very small on the contrary. These problems do not allow us to use common object detection models. Currently there are many computer vision algorithm that are designed using human centric photographs, But from the top view imagery taken vertically the objects of interest are small and fewer features mostly appearing flat and rectangular, certain objects closer to each other can also overlap. So detecting most of the objects from the birds eye view is a challenging task. Hence the work will be focusing on detecting multiple objects from those images using enhanced ResNet, FPN, FasterRCNN models thereby providing an effective surveillance for the UAV and extraction of road networks from aerial images has fundamental importance.