|Project Description: ||The development of Unmanned Aerial Vehicles (UAVs) is increasing, and the images taken by UAVs can supply considerable valuable information that affects our lives. It would be very useful to have a system that allowed UAVs to detect objects, or even provide users directly with semantic outputs or speech related to the images via the recording equipment onboard. Deep learning is a new field of machine learning that researchers in diverse fields are used to develop learning in neural networks. CUDA, NVIDIA’s parallel-computing architecture, increases the computing performance of the graphics-processing unit (GPU), and is used typically in deep learning to accelerate the processing speed. As a result, deep learning provides several advantages, in that it requires less training effort, and has a faster processing speed, greater accuracy, and the capability to process larger datasets. Recently, it has been used in image classification to aid in the recognition of specific objects in images. Using the deep learning framework, Caffe, the proposed system is able to identify objects in images taken by UAVs. Semantic outputs are generated to detect various objects, such as buildings, trees, etc., without additional translation. Because of its high processing speed, in future, we may be able to detect objects in real-time by implementing the proposed system on UAVs.