Transparent objects is a pipeline that can detect and estimate poses of transparent objects, given a point cloud model of an object. The pipeline is fully integrated into the Recognition Kitchen so usual training and detection from Object Recognition can be used with subsequent grasping. See ROS Quick Guide or Object Recognition for details how to use it.
Ilya Lysenkov, Victor Eruhimov, and Gary Bradski, “Recognition and Pose Estimation of Rigid Transparent Objects with a Kinect Sensor,” 2013 Robotics: Science and Systems Conference (RSS), 2013.
Ilya Lysenkov, and Vincent Rabaud, “Pose Estimation of Rigid Transparent Objects in Transparent Clutter”, 2013 IEEE International Conference on Robotics and Automation (ICRA), 2013.