Image processing Sight-enabled robots process large workspace with pinpoint accuracy

Related Vendors

MVTec Software GmbH

FAULHABER Drive Systems

Wuxi InfiMotion Technology Co., Ltd.

Automation specialist Akeoplus has developed a robot cell for the production of satellite panels at the aerospace company Thales Alenia Space. The special feature: Robots apply the components with very high precision on a workspace of 6 by 3 meters under cleanroom conditions. This is made possible by industrial image processing.

Increasing productivity, relieving humans of physically demanding tasks, and improving quality are the major advantages of robot-based automation. Industrial image processing (Machine Vision), the eye of production, ensures that robots reliably fulfill these promises.

A specialist in implementing robot-based automation solutions in conjunction with Machine Vision is the company Akeoplus. The company, headquartered in the French Château-Gaillard, is considered a pioneer in the field of robotics. Akeoplus was founded in 2006 and supports customers from various industries with applications for optimizing industrial robots. As part of its robotics solutions, Akeoplus uses modern image processing systems and advanced algorithms.

Gallery

A system that combines the strengths of robotics and industrial image processing was developed by Akeoplus in cooperation and on behalf of the Thales Alenia Space site in Cannes, southern France. Thales Alenia Space is a global space corporation that offers high-tech solutions for telecommunications, navigation, earth observation, environmental management, exploration, science, and orbital infrastructures. The engineers at Thales Alenia Space have developed an innovative adhesive process that can automate the placement of Optical Solar Reflectors (OSR), while also increasing thermal performance and weight. The experts at Akeoplus then worked with the Thales Alenia Space teams to develop a robot solution for a satellite panel manufacturing process.

In the system named "Solar", robots apply Optical Solar Reflectors (OSR) to satellite panels. The task in the development was to enable the robots to apply the OSR to the 6 by 3 meter panel with an accuracy of less than 0.2 millimeters under clean room conditions. The ratio of positioning accuracy to the size of the working area is 1:40,000, which posed a particular challenge in the implementation. To meet the requirements, Akeoplus relied on the Machine Vision software Mvtec Halcon, which supports inspection activities and alignment of the OSR.

Benefits of automation

Optical solar reflectors are used in space travel, for example to protect satellites in geostationary orbits from the high radiation levels there. A corresponding panel as an outer wall is equipped with up to 7,000 reflectors, with an OSR measuring 40 by 43 millimeters.

Our customer can now produce much faster, even though we have introduced additional steps for traceability and quality assurance.

Maxime Motisi, COO with Akeoplus

Until now, the OSR were applied to the panel purely manually. Two or more fitters lying on mats had the task of positioning the OSR components with their arms stretched downwards and ensuring their correct alignment.

"The production of large panels has increased, and working on such surfaces was not ideal for employees for ergonomic reasons. Thales Alenia Space is a major player in the field of space equipment and is constantly striving for excellent working conditions and the quality of its processes. That's why we decided together to automate the entire process. The robot cell was developed to speed up the process and thus increase productivity. Our customer can now produce much faster, even though we have introduced additional steps for traceability and quality assurance," explains Maxime Motisi, Chief Operating Officer and Project Manager of the Solar Robot Cell at Akeoplus.

Thanks to Machine Vision, two robots work autonomously

For the new automated production process, Akeoplus, together with Thales Alenia Space, developed the Solar robot cell. Inside it, there are two robots and a positioning table on which the satellite panels rest individually. For absolutely precise guidance, a laser tracker is responsible for the 6-axis robot arm, and for industrial image processing, three high-resolution 2D cameras are used on the hardware side. The Halcon image processing software is responsible for the quality check of the OSR, fine positioning, the human-machine interface, and visualization.

The manufacturing process proceeds as follows: The smaller and faster robot picks up an OSR from the tray and places it on a vacuum backlight under the first camera station. The camera takes a picture, which is examined by Halcon for fractures and scratches. In addition, the exact position of the OSR is determined and checked whether the component in size and shape also corresponds to the model provided for in the blueprint using the Machine Vision method "Shape Matching".

The position determination is so important because the robot picks up each optical solar reflector slightly differently. Therefore, the OSR is aligned using Machine Vision in order to transport it to the next production step with an accuracy of around 0.2 millimeters. If the component is "OK", it is placed under the nozzle of the adhesive dispenser. After the adhesive is applied, the smaller of the two robots picks up the OSR again and places it in a designated spot so that the larger robot can pick it up for further processing. The robot with the larger workspace now moves the component in front of the next image processing station, where the second camera takes images of the applied adhesive. The extracted image of the adhesive pattern is analyzed by Halcon. It checks whether the adhesive is evenly distributed and adheres to a very specific pattern that thermal engineers from Thales Alenia Space have developed and that must be reproduced exactly. This is followed by the precise alignment of the OSR on the robot's arm. This is the first step in being able to attach the reflector to the designated spot on the satellite panel with high accuracy. For this, the third high-resolution camera again takes pictures of the OSR. The alignment is carried out by Halcon's algorithms determining the spatial transformation between the outer edges of the OSR and the coordinate system of the gripper, which holds the OSR with a vacuum gripping system.

The corners of the OSR are determined by the intersections of the edges. From this a coordinate system can be derived in which the corner of the OSR is centered and aligned to the edges of the OSR. The coordinate system of the effector robot was determined in advance by a hand-eye calibration of the robot.

The determination of the coordinate systems is important for two reasons: First, for 3D alignment. It may be that the reflector is tilted on the gripper. Therefore, the corners of the OSR are matched with the effector, each corner being known and thus serving as a reference point. This allows even minimal tilts or shifts to be compensated for. Secondly, the corner of the OSR serves as a reference point for mounting on the panel.

However, since the 2.4 meter long gripper arm of the robot is not 100 percent accurate, but the reflectors must be attached with very high precision, the inaccuracies must be compensated. With the support of the laser tracker and the image processing software Halcon, Akeoplus has developed an iterative feedback process. This enables the robot to attach the OSR with the required accuracy at the exact spot on the 18 square meter panel.

Aligning the OSR on the panel with sub-millimeter accuracy

"We have been working with Mvtec for many years now. Therefore, we knew that with Halcon we have the powerful Machine Vision software that we need for the Solar Robot Cell," explains Maxime Motisi. Akeoplus uses the C# programming language to control its systems. With the help of the Hdevelop development environment integrated in Halcon, seamless process integration can be enabled, for example by loading the programming of the system control in Halcon and developing it further. Then, the Akeoplus Human-Machine Interface can represent each Halcon process as a function with the inputs and outputs to be used in the process sequence. This is helpful in order not to mix the work of the C# developers and the image processing expert: The image processing process takes place within a capsule, uses images and parameters as input, and returns results or values. Another important factor for the implementation was the speed and the ability to quickly develop or adapt powerful algorithms that can in turn be easily integrated into the "host system".

On the other hand, a powerful image processing software like Halcon offers technologies to perform, for example, three-dimensional position determinations or to detect errors in components. Specifically, objects can be reliably found with a higher accuracy than 1/20 pixel using matching methods, even if they are partially obscured. In addition, images can be suitably aligned. Subpixel accurate shape matching is one of the most important technologies. It enables the precise, robust and very fast finding of objects. In terms of quality control, in addition to checking for defects, determining dimensional accuracy is essential. It checks whether the components are within the defined tolerance ranges. Last but not least, industrial image processing enables subpixel accurate measurement of edges along lines or circular segments in less than a millisecond.

Application also serves as a showcase

The Solar Robot Cell is already in operation. "The implementation was not quite easy. Especially the work on and with the adhesive was challenging, but thanks to Thales Alenia Space's expertise in thermal adhesives, we were able to find solutions together. We can now be very pleased with the result, as the requirements in terms of quality, productivity and employee relief were all met," says Maxime Motisi. Specifically, this means that the cycle time for gluing and positioning an optical solar reflector has been reduced to less than 45 seconds and an OEE of >95 percent is achieved. As a result, a large panel can now be completely clad with solar reflectors by just one employee in five days. Previously, about two employees were busy with a panel for seven days. In addition, the work environment for employees has improved, as the tasks that took hours have been eliminated and contact with the chemical adhesive has been reduced. In addition, a new quality assurance process was successfully introduced, namely the inspection of the OSR for fractures and scratches.

In addition, the robot cell serves another purpose. Maxime Motisi explains: "With our Solar Robot Cell, not only can satellite panels be produced for actual use. The application is also a showcase project for Thales Alenia Space's robotics solutions, demonstrating their competencies and expertise. And we at Akeoplus are very proud of this use case. I would say that this case represents a summary of all our capabilities and knowledge, and it allows us to attract more customers for our innovative technologies."

For Akeoplus, the combination of robotics and industrial image processing has always been a necessity to meet requirements that have never been implemented before. "Our value lies in the connection of cutting-edge technologies with robots. Mvtec offers one of these key technologies and has already been a partner for more than eight years," concludes Maxime Motisi.

 (kip)

This article was first published on our sister brand 'elektrotechnikAUTOMATISIERUNG' (German Edition), Vogel Communications Group