How Robots Can See. The Power of Machine Vision
ABAGY Robotic Systems
Robotic Welding. Autonomous Robots. AI & machine vision-based software. Just click & weld.
Robotic welding has transformed the manufacturing industry by improving efficiency, productivity, and precision. However, the latest advancements in machine vision have taken this technology to a whole new level. In this article, we explore the remarkable capabilities of machine vision in robotic welding and its impact on enhancing safety, accuracy, and adaptability in welding processes.?
How Machine Vision Works
The integration of depth camera sensors has revolutionized machine vision in robotic welding. Unlike traditional cameras, 3D cameras have the ability to perceive objects not only along the X and Y axes but also along the Z axis.?
At ABAGY, we employ a range of machine vision devices tailored to specific tasks. Moreover, our approach allows for flexibility in the placement of these devices.
To illustrate the concept, let's take a look at a concrete example: we have mounted a compact and cost-effective snapshot scanner onto the robot's flange.
For a simple test, we randomly place a part in a cell. We also arbitrarily install clamps on the part. After that, we upload the 3D model of the part into the ABAGY software.?
Optimizing Scanning Process
The machine vision system powered by the snapshot scanner utilizes advanced algorithms to plan optimal scanning paths. The system finds a balance between the number of shots and data quality and can adapt to different angles and positions by moving the robot and using positioners.
Stitching Data for a Complete Picture
As each shot from the snapshot scanner captures a specific perspective, the collected data needs to be seamlessly stitched together to form a complete 3D model. Behind the scenes, robust algorithms work diligently to align and combine the captured data, ensuring a cohesive representation.
Matching 3D Model and Real Part
Upon obtaining data from machine vision regarding a real part within a robotic cell, ABAGY's advanced algorithms perform a matching process between the acquired data and the 3D model previously uploaded into the software. This enables the system to identify any deviations and subsequently propose the optimal robot path required to successfully carry out the welding task.
Identifying Fixtures Without 3D Model
Machine vision in robotic welding enables precise object localization. Additionally, one of the remarkable capabilities of our system is its ability to accurately identify fixtures without pre-existing 3D models. Unlike parts, fixtures often lack readily available 3D models from customers. To simplify the process, we automate the recreation of 3D models for fixtures using machine vision. These models are slightly expanded to ensure collision avoidance while maintaining operational efficiency.
Efficient Decision-Making Interface
The machine vision system provides operators with an intuitive interface that streamlines decision-making. Processed data is represented through green and red points in the point cloud visualization. By following the guideline of "more green and less red," operators can swiftly assess whether the welding process can proceed as per the task requirements.
领英推荐
Challenging the System
In order to challenge the capabilities of the machine vision system, we intentionally introduced various obstacles and objects into the welding cell. To add clarity and a touch of excitement, we decided to use our engineer's iPhone. Placing the phone directly on the part, right on the weld, we wanted to see how the system would respond.
However, the system proved to be quite resilient and didn't fall for the trick. After scanning, it informed us that welding was not possible due to the obstruction caused by the phone. We made adjustments to the welding parameters, ensuring that the phone no longer interfered with the process. With the new scan data provided to the system, we were now ready to proceed with a dry run.
To make the task even more challenging, we decided to introduce a few additional objects into the welding cell. Taking whatever we could find around us, we included a coffee mug and a toy from one of our engineer's children. After the test, the Teddy Bear, mug, and phone remained unharmed and intact.
Watch the video on our YouTube channel to see how the tests unfolded and gain a visual understanding of the results.?
In our upcoming newsletter, we are excited to continue the discussion on machine vision. In the next issue, we will delve into the topic of different types of machine vision devices, providing insights into their characteristics, functionality, and points of differentiation. Stay tuned for an exploration of the topic of Autonomous Robots.?
Author:?Kate Degai, CMO at?ABAGY Robotic Systems
Follow us on?LinkedIn?to stay in the loop with all our latest updates and discoveries.
Best wishes,
ABAGY Robotic Systems?team.