Using 3D imagery and robotic helicopters, the Navy plans to scan the seas for pirates.
The plan is to equip a drone helicopter, called Fire Scout, with the cameras and the LADAR sensor in a system the Navy calls Multi-Mode Sensor Seeker (MMSS).
Once images are collected, software can sift through the data, comparing collected images to images of reference vessels, to help narrow in on potential problems. Without some sort of automation to process the images, sailors will be overwhelmed with data.
"The 3D data gives you a leg up on target identification," Dean Cook, principal investigator for the MMSS program at Naval Air Warfare Center Weapons Division, said in a statement. "Infrared and visible cameras produce 2D pictures, and objects in them can be difficult to automatically identify. With LADAR data, each pixel corresponds to a 3D point in space, so the automatic target recognition algorithm can calculate the dimensions of an object and compare them to those in a database."
The Office of Naval Research said the software algorithms have been tested from shore-based systems on vessels at sea. Now the Navy plans to test the software on a manned helicopter.
The system is a good example of how robotics and machine learning is opening up new applications, including security and surveillance. LADAR systems send lasers beams onto an objects and process the signal echoed from the targets to create a 3D image. The Office of Naval Research said the lasers are "eye safe."
As autonomous vehicles, such as the drone Fire Scout, become more widely used by the U.S. military, the need for robotic and machine-learning software systems has gone up. The Navy last week opened its Autonomous Systems Research lab to test a range of robotic systems, including handheld quadcopters and a mobile robot designed to extinguish onboard fires.