Validation of Airborne Bathymetry Using Long-Short Term Memory Neural Network

Drones are adaptable technologies that can be used in several applications. They are especially effective in monitoring and surveying tasks due to the low cost and time embedded in such operations. Unmanned Aerial surveys can be repeated with minimal abortive cost and do not require special setup. The ability to carry wide spectrum of airborne sensors render these platforms optimal and compatible with different flight missions and real-time scenarios. This paper carries out a study on bathymetric technique. Datasets obtained by real airborne bathymetry are used to train an LSTM neural network that will be implemented as quality control for the fathometer reading prior to further processing. The extrapolated values will serve as an orienteer for flight mission success depending on auto and partial correlation of the dataset.