Most industrial applications of computer vision can be categorized into two groups. They are (1) visual inspection and (2) machine parts recognition. There are several review articles for automatic visual inspection [1,30,31]. This paper gives a brief review of robot vision system for machine part recognition. A robot vision system for machine parts recognition contains four sub-systems: (1) sensing, (2) segmentation, (3) description, and (4) recognition. A block diagram of such a system is shown in Fig. 1.
Sensing is the mechanism by which robots receive information from the world around them. Robot sensing, and in particular, robot vision is the basis for providing sensory inputs to a smart robot. Devices that make robot vision possible are typically classified as electro-optical sensors. Thus a critical review of robot vision is essentially a review of current and predicted capabilities for imaging electro-optical sensors. The critical element in an electro-optical sensor is the detector and most detectors are semiconductor devices. Thus we conclude that robot vision is closely tied to the state of the art in the development of photosensitive semiconductor materials and in the fabrication of large scale arrays of detector elements. This paper begins with a review of the fundamental principles that are important to electro-optical sensors. It then presents a brief tutorial on parameters for quantitatively evaluating the performance of detectors and electro-optical sensors. Selected state of the art sensor systems are next analyzed to show expected performance. Finally a review is given of detector technology as related to robot vision including a comparison of materials, array sizes, performance and cost.
The use of industrial robots in a non-ideal environment requires sensors. A versatile general-purpose sensor is a machine vision system. Rather than attempting to review the capabilities of commercially available machine vision hardware, this paper examines the application of both binary and gray scale techniques to the sensory tasks associated with robot part acquisition, part inspection, and part reorientation. The utility of hybrid (mixed binary and gray scale) techniques in robot vision is illustrated.
The problem of recognizing multiple objects in a highly cluttered background in the face of geometrical object distortions is addressed in this paper. A correlation architecture using a matched spatial filter of a synthetic discriminant function is employed to achieve the required performance. Synthesis of the synthetic discriminant function is discussed as is the initial performance obtained in the face of noise. Initial remarks are advanced on various methods to select the training set of images to use in this algorithm.
The purpose of robot arm control is to maintain a prescribed motion for the manipulator along a desired trajectory by applying corrective compensation torques to the actuators to adjust for any deviations of the manipulator from the trajectory. This paper presents various control methods for industrial robots. It begins with the discussion of various dynamic models for manipulators and covers several existing control methods from simple servomechanism to advanced controls such as adaptive control with identification algorithm.
Components of robot motion control are described. Deficiencies of existing methods of motion control are dis-cussed, and possible solutions are indicated. State-of-the-art, as well as trends in the development of components of robot motion control, are traced.
Teleoperated manipulators have been used for many years to perform tasks within hazardous environments. There are many similarities between teleoperators and robots, and these likenesses are delineated and reviewed. Applications and development activities in teleoperated systems are summarized on a worldwide basis. Teleopeator developments are examined for outer space, under water, and other hazardous environments. The unification of a robot and a teleoperator system into an autonomous, flexible machine is envisioned as the goal of future "telerobotic" research.
Robots are attracting increased attention in the industrial productivity crisis. As one significant approach for this nation to maintain technological leadership, the need for robot vision has become critical. The "blind" robot, while occupying an economical niche at present is severely limited and job specific, being only one step up from the numerical controlled machines. To successfully satisfy robot vision requirements a three dimensional representation of a real scene must be provided. Several image acquistion techniques are discussed with more emphasis on the laser radar type instruments. The autonomous vehicle is also discussed as a robot form, and the requirements for these applications are considered. The total computer vision system requirement is reviewed with some discussion of the major techniques in the literature for three dimensional scene analysis.
1 By 1990, over 100,000 industrial robots will be operating in the United States. This dramatic increase in the use of industrial robots will require robot vision systems so that the machines can respond and adapt to their environment, i.e., be intelligent. These advances can lead to flexible manufacturing systems and to automated factories. The purpose of this paper is to provide a review of machine vision techniques, with special emphasis on three dimensional perception and methods for non-contact measurement of the coordinates or surface normals of objects using stereo, shading, and projection techniques. These methods include two dimensional model matching, shape from shading, stereo, and single image stereo. Surface fitting methods are then considered which permit one to determine which mathematical model best fits a simple surface. These techniques for three dimensional measurement and recognition are essential for machine perception for intelligent robot applications.