Break-Through Visual Pattern Recognition for Robotics and Automation

Visual recognition of patterns is a basic capability of most advanced species in nature. A big percentage of the human brain is devoted to visual processing with a substantial portion of this area used for pattern recognition. Visual pattern recognition enables a variety of tasks such as object and target recognition, navigation, grasping and manipulation, among others. Advances in camera technology have dramatically reduced the cost of cameras making them the sensor of choice for robotics and automation. Vision provides a variety of cues about the environment (motion, color, shape, etc.) with a single sensor. Visual pattern recognition solves some fundamental problems in computer vision: correspondence, pose estimation, and structure from motion. Therefore, we consider visual pattern recognition to be an important, cost-efficient primitive for robotics and automation systems. Recent advances in computer vision have given rise to a robust and invariant visual pattern recognition technology that is based on extracting a set of characteristic features from an image. Such features are obtained with the Scale Invariant Feature Transform (SIFT) [3, 4] paradigm and represent the variations in brightness of the image around the point of interest. Recognition performed with this types of features has been shown to be quite robust in realistic settings. This paper describes the application of this particular visual pattern recognition technology to a variety of robotics applications: object recognition, navigation, manipulation, human-machine interaction. The following sections describe the technology in more detail and presents a business case for visual pattern recognition in the field of robotics and automation.