Machine Vision Archives - Robotics Academy

Precise measurement using machine vision

The Robotics Academy was commissioned to investigate further the potential of machine vision for Neorem Magnets’ needs. Atte Ali-Hokka and Juha Aalto were involved in the project in the Robotics Academy team. The project was directed by Joonas Kortelainen. The task of the team was to investigate the possibility to measure the object delivered by the company with a machine vision camera in the Satakunta University of Applied Sciences’ RoboAI collaborative laboratory.

The customer wanted to study the technical dimensions and quality of the object as accurately and reliably as possible. The project used an IDS camera that captures the subject in very high resolution. The team implemented the program in Halcon development environment that meets the customer’s requirements. The group had no previous experience with machine vision, but the Academy students were happy to take up the challenge.

The client was particularly interested in exploring the possibilities of measuring and examining the area of the object and its formal correctness compared to the technical drawings provided by the company. The students had no previous experience in creating such a program so students went on to study the measurement of the object by creating a program that can measure the length, angles of sides and perimeter of the object.

As the team learned about the new object an image was created in the real world coordinate system allowing the size of the object to be measured in millimeters. This development step opened the doors for the working group to a whole new way of measuring, as it was now possible to compare the image to the technical drawings provided by the customer to the working group.

Summary:

The customer seemed satisfied with the results achieved by the working group. The students and the client felt that both parties had benefited from the project. For students, this benefit is reflected in the amount of learning that students gained from working on the project with a great deal of machine vision, as well as camera, lighting, and programming techniques. The Robotics Academy would like to thank Neorem Magnets for taking the opportunity to explore their expertise and provide an interesting project.

Machine vision as an industrial tool

Our client, Boliden Harjavalta’s main products are copper, nickel, gold and silver. Boliden’s nickel smelter is the only one in Western Europe. Boliden produces high quality nickel matte using the latest technology in the industry.

The Robotics Academy students were commissioned to explore the possibilities of using machine vision in the Boliden process. Boliden was interested in the concentration of various components of the slag after the smelting process and hoped to use machine vision to simplify and speed up the process.

A machine vision system is one where computer imaging is used for industrial purposes. The system consists of a light source, an object to be photographed, a camera, a computer, and an image processing program that interprets the image automatically. Machine vision systems perform precisely pre-programmed tasks and are primarily used when optical inspection should be fast, accurate, round the clock and consistently repeatable. Machine vision can be used to perform tasks that are impossible for human vision by using wavelengths that cannot be detected by the human eye. In this assignment, a conveyor belt cared slag under the system to identify the customer-defined concentrations.

When using a smart camera, all image processing and counting is done in the camera itself. We chose to use Cognex’s In-Sight application, but the slowness of this camera was a disadvantage because the items had to be pictured on a fast conveyor belt. Due to scheduling concerns, Robotics Academy students were not able to continue their research further, however, it is possible to continue the project with another software and a different camera solution in the future. We recommend considering Halco software, which could handle all intellectual observation and computing, as well as speed up image processing.

Summary

In our project, we learned how to take advantage of different camera and light options. The overall benefits were mutual. The students’ gained knowledge and were able to give the customer the information they sought. With the help of additional instruction, we can better utilize machine vision in future projects.

Automated beverage bottle opener

Satakunta University of Applied Sciences has a Universal Robots 5 (UR5) robot arm, which is utilized for educational purposes by students as well as for school demonstrations. A demonstration that has proven popular at public events includes UR5 dispensing drinks. Through code, the robot is directed to use its machine vision to take a bottle from a bottle rack, pour the drink into the cups and drop the bottle into the trash. One major issue with this demonstration is that the bottles need to be manually opened. The number of drinks served at these public events range from tens to hundreds, depending on the audience, so it made more sense to use a machine for opening bottles.

This case was given to the students of the Robotics Academy to solve. After examining and analyzing the current situation, they discovered that the robot’s own grip strength was not enough to be able to spin the cap open and separate it from the bottle. The group responsible for this case mapped possible solutions, taking programming, mechanical and functional needs into account.

The biggest challenge was designing a gripper/opening head, which would suit the majority of bottle types and would also work for the robot, in terms of possible features. From the mechanical implementation point of view, the multifunctionality of the opening mechanism presented a challenge. The opening head had to rotate to remove the cap and then the cap had to be removed from the device before the next beverage bottle was brought to the machine. Used caps and bottles had to be deposited in the trash to make the presentation table ready for the next bottle.

The body of the device is constructed of an aluminum profile and the cage was designed from plexiglass. SolidWorks -3D design software was used to design mechanical parts, each track and gear carefully modelled to fit together. When the models were ready, they were printed on the school’s 3D printers. We chose biodegradable and inexpensive PLA plastic. In this way, making the new parts was as easy with the ability to print more parts, if needed.

For programming the opener, we chose Ardo’s Uno. It is a small microcomputer that is programmed in C ++ programming language.

Summary

The working solution: The robot takes the bottle under the opening head and the attachment removes the cap. The necessary equipment was made to place the removed cap in the garbage. We tested the operation of the device, and after the adjustments we got it to work consistently.

Collaborative robot in assembly work

Robotics Academy was commissioned by Oras Group to investigate the robot’s suitability for assembly work. The goal was to automate the five-part line using a assembly robot. ABB’s two-handed YuMi® robot, developed for the assembly of small parts, was selected. Robot has flexible position-changing hands, camera-based positioning and accurate control.

Academy’s students were responsible for programming and simulating the robot, using machine vision, and designing and implementing parts of the robot’s tools, ie. jigs, palettes and levels, by 3D printing. Students in the project have a wide-ranging understanding of robot activity and constraints. The most important goal was achieved – the assembly task was accomplished with a robot.