Intelligent agricultural automation using UAV based computer vision and deep learning algorithms

The use of Unmanned Aerial Vehicles (UAVs) in agriculture has continued to grow at a phenomenal pace in the past few years due to the maturation and affordability of associated technologies, both in software and hardware. Among the many open problems, one which this research direction will aim to address is how to achieve real-time onboard execution of computer vision and machine learning algorithms on UAV aerial imagery that can produce the level of accuracy required for full automation. To achieve this goal, it will not only require drastic improvements on state-of-the-art algorithms and models, but it will also demand innovations across the entire processing pipeline, including hardware acceleration, memory management, data reduction, parallelization of algorithms, software architecture, only to name a few.