An integrated low-cost system for object detection in underwater environments

We propose a novel low-cost integrated system prototype able to recognize objects/lifeforms in underwater environments. The has been applied detect unexploded ordnance materials shallow waters. Indeed, small and agile remotely controlled vehicles with cameras can be used bombs waters, more effectively freely than complex, costly heavy equipment, requiring several human operators support boats. Moreover, visual techniques easily combined the traditional use of magnetometers scanning imaging sonars, improve effectiveness survey. proposed adapted other scenarios (e.g., archeology or inspection pipelines implants), by simply replacing Convolutional Neural Network devoted identification task. As final outcome our work we provide large dataset images explosive materials: it compare different on common basis.