No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 Taking Control Of The Multitude Of Varieties In Automotive Engineering: Robot Vision Increases Flexibility In Production The winner is: 3D with one camera For almost any requirements in modern manufacturing, there are optimized image processing solutions offered by ISRA VISION. The vision systems have proven themselves to be indispensable tools particularly with regard to robot guidance and positioning. With the MONO3D systems, new areas of application have been developed, that make it possible to use only one camera to precisely determine all six degrees of freedom. As a result, among other things, the growing variety of components being used in automotive design can be handled with ease and economic solutions can be implemented for robot applications, which can be seen with two examples at both General Motors and Daimler Chrysler in the United States. Without image processing technology, robots are limited in their capabilities. Not until the robot's vision system has been able to exactly identify the component and its position is a defect-free operation possible. The so-called robot vision allows higher production rates, it increases and safeguards the quality and decreases the costs of production. With this aspect alone, an investment in an image processing system can pay for itself often within just a few months. Especially in the automotive industry, the number of pieces manufactured is increasing while placing more pressure on cost management. At the same time the number of models is also growing Page 1 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 significantly. This can be seen when using Audi as an example: While in 1996 the company manufactured close to 500,000 vehicles, in 2005 that number was considerably higher at 800,000 units. For 2015, the company even expects to be manufacturing 1.4 million vehicles. At the same time the choice of products is expanding dramatically. In 1994, Audi manufactured seven models. Eleven years later, as a car buyer, you had 22 models to choose from and for 2015, according to Audi sources, 40 different models are expected. The range of products is continuously being expanded with new model series, customized and niche vehicles. This development, which other automotive manufacturers are experiencing in a similar manner, requires highly flexible production processes, which can only be accomplished using robots in combination with an intelligent image processing solution in order to precisely recognize a growing number of components. Optimum solution for any demand Industrial image processing gives the robots intelligent "eyes" to see with. The image-processing specialist, ISRA VISION, provides imageprocessing systems based on a number of different technologies for every demand, which assure better productivity and efficiency in every processing stage in the production chain of the automotive industry. Even 2D systems can be used to determine positions, identify types and components on the conveyor belts. An additional step is made possible with 2 ½ D technology, which calculates the distance of the captured object as a "half" third dimension from one single camera position. Using this technology, robust and economically efficient solutions can be realized during palletizing and depalletizing or for sorting containers. Page 2 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 Three-dimensional image processing helped robot vision to take a substantial leap forward. ISRA provides a comprehensive program of "real" 3D methods with which all tasks can be carried out. The 3D stereo sensor determines positions and measures objects in all six degrees of freedom. It can be used in a single position or be mounted to the robot arm as a portable device. The 3D photogrammetry systems are mostly used when very large or multiple parts are to be captured at high-resolution levels. 3D robot vision uses the information available from numerous cameras to do this. The position of entire automotive bodies can be captured precisely at lightning speeds, for example in paint finishing systems. In areas when the quality and the efficiency of highly automated robots and handling systems need to be increased through fast and precise position definition, the 3D stereo gauging technology can be used. 3D form matching combines robot vision and inline gauging technology using multiline project sensors and integrated surface illumination. Using one camera to see three dimensionally Currently another innovation is producing a complete breakthrough for robot vision, even from an economic point of view. The MONO3D System, which ISRA VISION introduced to the market about two years ago, is highly efficient, space saving and, additionally, cost effective. The vision system is based on a newly developed software program and only needs one camera in order to be able to gauge numerous points from one position in a 3D camera field with precision and speed, Page 3 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 which is otherwise only possible using expensive multi-camera systems or lasers. The system generates 3D position data at high cycle times by fixing the position of three or more features of an object, by then defining the spatial relationships between these features and comparing the results with a stored 3D CAD model. All features of an object that are recognizable in the camera image can be used for measuring purposes. This way, all six degrees of freedom (position and orientation) of any point within the camera's field of view can be determined precisely. This approach not only drastically reduces costs for materials, but also the expense for installing, calibrating and operating the system. In the event of geometric disadvantages, numerous camera positions can be combined with each other. The system works without any problems even at high production speeds. Because of its clearly arranged operator interface, it can be calibrated and installed easily. In use at General Motors and Daimler Chrysler The MONO3D system makes it possible to expand the use of robots in many more areas of applications that have, up to now, been unavailable for economic reasons. The automotive industry and their suppliers, which is under a great amount of cost pressure, quickly recognized the benefits of this system and uses MONO3D in many applications, for example in a General Motors engine casting facility in Saginaw, Michigan in the United States. In the casting facility located in the middle of the US automotive industry just northwest of Detroit, castings for various engines are manufactured using sand-casting Page 4 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 methods. In the process, compacted forms made of bound sand are used. These sand castings absorb the molten metal and complete engine molded parts are the result. The sand castings are stored in the interim on shelves. From there they are removed by robots and placed on conveyor belts so that they can later be assembled to a complete mold negative. In order to be able to determine the exact position of the correct form the mono 3D system is installed on the arm of the ABB robot above the gripper device. The robot navigates to this defined position in the shelf by first positioning downward. In the process, the highest level of precision is required so that the gripper correctly picks up the delicate sand castings. After that, the robot arm with the gripper devise navigates along the shelf and removes the sand casting, which is attached to the gripper using a pneumatic module. Mono3D systems can also be used to locate the form elements on the conveyor belt. All in all, 40 Mono3D systems are in use in the engine casting facility in Saginaw. Simple application that is convincing Even Comau Pico, one of the leading US integrators of automation systems for the automotive industry with headquarters in Southfield, Michigan is convinced of the economic efficiency and effectiveness of the Mono3D systems. The company that was founded in the US in 1939 and was taken over seven years ago by the Comau Group, uses these systems for Daimler Chrysler in numerous production facilities across the US. Especially the system's simplicity was a significant factor for this decision. Page 5 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 The MONO3D system is used for example at the Daimler Chrysler plants in body shell applications for depaletizing pressed sheet metal components. Using robots, it was possible to automate this unloading process. The difficulty of this task is that the sheet metal components are placed on the floor in containers and are used for a number of different applications so that they have no standard form. In this case, MONO3D systems are applied that use new miniature cameras and compact high-performance LED spots that can easily be integrated into the robot's gripper system. Intelligent image processing systems are the key to more efficiency and higher levels of productivity in the automotive industry. ISRA VISION includes robot vision as one of its core competences for more than 20 years and offers a complete line of image processing systems for robot guidance. MONO3D systems particularly make it possible to offer very cost effective image processing solutions that stand to be compared to any other system. Because of its global focus and infrastructure, the company is available in all relevant markets as a partner for vision solutions oriented perfectly to any specifications. . Page 6 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 Images MONO3D in use during automated door assembly (MONO3D_1.jpg) High processing speeds using state-of-the-art robot vision (MONO3D_2.jpg) Page 7 No. 132, 12.06.2007 Walter Meyer wmeyer@isravision.com ++49 (6151) 948-172 Complicated and costly mechanical separation devices are no longer necessary when "seeing" robots" are in use (MONO3D_3.jpg) Miniature cameras and illumination devices integrated into a complex gripper tool (MONO3D_4.jpg) Page 8