 This paper proposes a new approach to using Edge Computing in Drones to enable the processing of extensive AI tasks on-board UAVs for remote sensing. It introduces Aero, a UAV brain system with on-board AI capability using GPU-enabled Edge devices. Aero is a novel multi-stage deep learning module that combines object detection, YOLO-V4 and YOLO-V7, and tracking, deep sort, with TensorRT accelerators to capture objects of interest with high accuracy and transmit data to the cloud in real time without redundancy. Experiments showed a reduced false positive rate, 0.7%, and a low percentage of tracking identity switches, 1.6%. Additionally, the system achieved an average inference speed of 15.5 FPS on a Jetson Xavier AGX Edge device. This paper demonstrates how Edge Computing can be used to improve the performance of unmanned aerial vehicles in remote sensing applications. This article was authored by Anise Kuba, Adele Ammar, Mohamed Abdelkader, and others. We are article.tv, links in the description below.