RVT’s patent portfolio covers technology that has been developed and enhanced for over twenty years as well as new technology currently under development. eVf, or eVision Factory, is the brand name of the software that RVT distributes. It allows users to create scalable and robust vision guidance systems for industrial robots. Our six current patents are:
- Method and Apparatus for Single-Camera 3D Vision Guided Robotics (US 6,816, 755 B2)
- Method and Apparatus for Single Image3D Vision Guided Robotics (US 8,095,237 B2)
- System and Method of Three-Dimensional Pose Estimation (US 7,957,583 B2)
- System and Method of Determining Object Pose (US 8,437,535 B2)
- Method and Apparatus for Machine Vision (US 7,336,814 B2)
- Methods and Apparatus to Facilitate Operations in Image Based Systems (US 8,559,699 B2)
A component module of eVF is our Single Camera 3D (SC3D) technology. SC3D uses one camera to calculate vision information in six degrees of freedom. The first patent of its kind and on the above list
protects this technology. The first patent covers picking six [some times as few as four] unique features
on a part to train the vision system to recognize so that an accurate 3D position may be calculated for
the robot to go and grab a part.
The second patent covers the technology that once our vision system has been trained with a master
reference image, the part could be located anywhere in the working envelope of the robot and our
system can locate and direct/guide the robot where to go to pick the part up. The third patent
reference above covers how we train a workspace using multiple 2D images and translates that into a
3D object pose.
eVF contains validation tools that are built into the software framework. These tools are unique to RVT
and include Auto Calibration (AutoCal), Automatic System Training (AutoTrain), and Automatic System
Testing (AccuTest). The fifth patent listed protects all of the equations behind these tools.
Random Bin Picking (RBP- trademark of RVT) is an add-on module to eVF that allows a robot to locate
and remove parts from a bin. These parts can be randomly placed and completely unstructured. This
sixth patent on the list above one covers technology that we use to ensure that a part will be able to be
picked from the bin until it has been completely emptied.
Two patents were filed in 2014 that cover new developments that our scientists
have been busy working on. One is for finding features on a part automatically. In our version of eVF to
date, the user has had to specify which features on a part should be used in the master image for
reference. This new patent will find features on the part automatically and our software will be smart
enough to know which features on a part are good. This patent also covers automatically controlling the
lighting conditions in a work cell. Over the course of a day, the lighting conditions may change
drastically from when the vision system was setup. This new technology is able to control the lighting
automatically using feedback from the camera. Both enhancements significantly reduce costs for the
The second new patent focuses on humans and robots sharing a work cell, a module we call Safety Net.
Currently, expensive light screens and other sensors are used to protect workers from accidentally
breaching a robot’s working envelope. This new technology uses a camera-based approach and is able
to detect an object coming into contact with a robot. The user interface allows an operator to designate
zones where the robot would either slow down or stop, based upon where the object is at with respect
to the designated zones. RVT will be filing for a ninth patent in the fourth quarter 2014 that will combine
today’s advanced computing power with robot communications and 3D vision software. Game-changer.