An Autonomous Grasping System for EOD Robot in Weak Light Environment

In order to solve the problem that the autonomous grasping system of EOD robot based on binocular vision or RGB-D depth camera can not function effectively in the weak light environment without visible light auxiliary illumination, which limits the application scene, a two-stage target detection, positioning and motion guidance approach for autonomous grasping of EOD robot has been proposed, which is based on fusion of monocular visual feedback with geometry ranging and laser ranging information respectively. With a prototype system, the process of autonomous grasping of 57mm projectile targets under the conditions of indoor normal lighting and no visible lighting at night has been verified. The test results show that the autonomous grasping system can complete the autonomous search, detection and grasping of typical ammunition targets in a certain range without visible light illumination.