Summarizing your 3D options

One of my takeaways from Automate was that 3D is no longer a novelty, at least not as regards robot guidance. Pretty much every robot integrator with a Fanuc on their stand also had 3D system in place. These all used the same DLP approach where a series of lines are projected in quick succession onto the target surface.

Andy Wilson summarizes this, and all the other options, in “Choosing a 3D vision system for automated robotics applications” which was published on the Vision Systems Design website back in December 2014. It’s a very comprehensive and I don’t think he omits a single camera-based method. (He doesn’t mention LIDAR or time-of-flight methods.)

If you have any interest in robot guidance I’d encourage you to read his article. Even if you’re more interested in part inspection, as I am, it’s an excellent primer. Ever helpful, Andy includes links to the many camera and illumination companies with something to offer the 3D imaging world. It’s something you might want to bookmark.

Vision guided robots no longer remarkable?

You know a technology has passed some kind of acceptance threshold when it barely merits a mention. Such is the case with this application story, posted on the website of gauging specialist Renishaw and since reproduced on Designnews.com.

“Equator™ programmable gauges help create the ultimate automation cell for bearing machining and parts sorting” describes a number of systems that a Texas-based machining house is installing. The most interesting, to me anyway, is the last one. The article mentions, “…a cell that consists of two Equators, a FANUC LRMate 200iC 6-axis robot, multiple lanes of low-profile conveyor, a FANUC iR Vision system and an ATI quick toolchanger for the robot's end-effectors. The vision system tells the Equator what part number is being presented and what measurement program to run. Good parts are subsequently placed on the appropriate conveyor, and bad parts are placed on a scrap conveyor.”

The photo below shows the system. Notice how there seems to be no fixturing to locate the parts for robot pickup? My assumption is that, in addition to part recognition, the FANUC vision system is also figuring out where the part is, and sending the robot to those coordinates.

If I’m correct, this is a very cool cell. If the developers ever read this, I’d love to see a video of it in operation.

Vision guidance movies

When I’m short of things to write about I dip into YouTube, and today is such a day.

This poppadum picking and packing system amused me. Not only is it high speed with vision guidance, but it’s an interesting gripper – presumably vacuum. I also liked the tongue-twister aspects – such a pity it wasn’t developed by PPT.
 

 

It’s also interesting to note that one robot isn’t fast enough to get all the poppadums, necessitating the installation of additional systems.

I’m pretty sure the robots used in the system are ABB FlexPickers. These are particularly cool, so I thought you might be interested in learning more about them. The movie below demonstrates how the system is calibrated – an aspect of VGR often overlooked – and of course, emphasizes how easy it is to set up. Well I’m sure it’s easy if you know what you’re doing!

 

An advance in vision guided robotics

Georgia Tech is doing some interesting work on visual servoing. They’re using a time-of-flight camera to dramatically reduce the time it takes to move a robot end effector through a sequence of steps to complete a task.

If you watch the video below you’ll learn they see this as useful in bomb disposal and surgery. I can see it going further. Currently we teach a robot a task by driving it to a position and saving those coordinates. But how about showing the robot the part it needs pick up, and then letting it find that part in 3D space? I already have an application waiting for such a capability.

VGR movies

If you enjoy watching robots do the random bin picking thing, Danish vision and robotics specialist Scape Technologies has a treat for you. A host of movies, (also on YouTube,) show their single camera vision system, (they snap two images from different poses and do some stereo work to calculate a 3D map,) combined with some clever gripper designs. It all looks very impressive, so allow me to pour a little cold water on their parade.

First off, none of the movies show the bins being completely emptied. In fact they only seem to pick off the upper layers of pretty full bins. I’d be more impressed if they demonstrated picking the last few pieces from the bottom.

Second, it all looks rather slow. Sure, a single camera approach cuts the investment slightly, but at the expense of cycle time. I’d want to just strap on a second camera and save the two seconds that’s spent reorienting and acquiring.

Now in fairness, what they’re doing looks good. I’m just not sure it’s any better than what I’ve seen from other companies. And I think everyone struggles to completely empty the bin. In other words, I don’t believe bin picking is truly ready for prime-time.

VGR primer

Go to any automation show and you’ll see robots guided by machine vision. They make it look so easy, although you know it’s not. But how do you go about getting up to speed with vision guidance?

Well you start by reading “The gift of sight: A four-step approach to setting up a machine vision system for robot-based manufacturing” This straightforward article, published on automationmag.com, November 17^th, 2011, discusses the main points to consider, in jargon-free English.

A particular strength of the article is how it covers robot-to-vision calibration. This is something I don’t think I’ve seen covered anywhere else, yet it’s essential (along with lighting,) to a successful implementation.

Read and learn.

Why Scorpion?

Staying with the vision-guided-robotics (VGR) theme of my last post, the marketing machine that is Thor Vollset recently drew my attention to his work on 3D bin picking. This was in support of the Autocast project, which is something I’ve been following closely.

In a blog posting on September 13^th Thor shows a video of a robot emptying a bin of castings. (“This is Autocast 3D Bin Picking for the Casting Industry.) We could nit-pick about the extent to which the castings were in a known and ordered arrangement for the video, but in the main it’s quite impressive.

But here’s my question: the last time I visited the Robots & Vision Show in Chicago every robot company had vision guidance. It seems that all of them, or at least the majors like my corporate robotics partner, is already in bed with a vision company. This means that if I need robotic bin picking I call up ABB, Fanuc or whoever and they quote me a turnkey package. I’m not sure why I should disregard their standard offering and instead go with Scorpion.

Thor, I know you’re a regular reader so perhaps you’d like to comment. (Make it a good one and perhaps you’ll get your own guest blogger spot!) Please tell us what Scorpion provides that I can’t get from my robot supplier’s existing partner.

Robot primer

Robots and vision systems go together like strawberries and cream, so today I’m going to deviate from my strict focus on machine vision by directing you to what I found a useful primer on the different types of robot available. This is “Robotic Automation for Solar Cell Manufacturing,” found in the June 2010 edition of Photonics Spectra. (Yes, that’s the same issue I was reading yesterday. I had a long flight.)

The article is written by Rush LaSelle of Adept Technology (www.adept.com ) but it struck me as a very balanced summary of the different options available.

Rush describes these as Cartesian (your basic X, Y, Z translation,) SCARA, articulated (the type most of us picture when we talk about robots,) and the most recent evolution of the species, the delta or parallel kinematic robot. If you’ve never seen one of these latter beasts at work, you’ve missed a treat. You can probably find a few movies on YouTube; just search for “Quattro,” the Adept delta robot, or “Flexpicker,” the competing product from ABB.

Vision guidance is being used in more and more robot applications, but the area in which it seems to make most is those involving delta robots. These are used chiefly for picking randomly placed parts from a fast-moving conveyor and I don’t know how the task would be done without vision.

If you work in machine vision, and especially if you’re at the sharp (application engineering,) end of the business, it would be prudent to understand the merits of the different technologies, so use the link I gave you.

An unsexy vision-guided-robotics (VGR) application

Here’s an interesting machine vision story from the OptioIQ web site (that’s the online home of Vision Systems Design magazine, if you didn’t know.) Under the heading “Plumbing Parts Assembly” Andy Wilson (come on Andy, is that really the best title you could come up with? Working to a deadline?) describes, in glowing detail, a system recently installed at a UK maker of pipe fittings.

What I like about this application is the emphasis on keeping it simple. Rather than try to pick parts from a bin, the integrator opted to feed them out in single file on a conveyor. This of course makes it much easier to find a candidate for the robot to pick. (It helps that the pipe fittings, at least those shown in the embedded video,) have only one natural pose.)

The lighting is however something of a puzzle to me; I can’t help thinking there is a better way – maybe laser triangulation, or perhaps stereo cameras – but then again, it seems the integrator was striving for robustness, not sexiness.

Tips and tricks for vision-guided-robotics

Vision guidance is one of the up-and-coming machine vision application areas. Everyone wants to do it because there’s no shortage of tasks to automate. The problem though is that few people have real been-there, done-that, VGR experience. Sure, vision guidance is all over the place at the trade shows, but real, practical, working applications are rather thin on the ground.

So if a VGR project is looming on your horizon you might like to do a little reading. One of my favorite places for machine vision application info is Quality magazine, and specifically their “Vision & Sensors” supplement.

The July 2010 issue has a couple of good articles on the subject of VGR. There’s “The Power of a New Dimension” by Wolfgang Eckstein and Lutz Kreutzer of MVTec which deals with approaches to 3D vision (essential if you want to locate a casting in a bin.) And then there’s “When Robots See Clearly” by Kevin Taylor of ISRA Vision Systems. This gets more into the options for vision guidance, but taken together, these articles could provide a good starting point for your VGR feasibility work.

Insight-full vision-guided-robotics application

There’s an interesting machine vision application described in the May 2010 issue of Assembly: “Robots Light the Way to Faster Assembly”.

This covers the use of a Cognex InSight with an ABB FlexPicker that assembles LED lights. The InSight is used to determine the orientation of lenses that are picked and placed over the LEDs by the robot.

What’s particularly interesting is the speed the system runs at. These FlexPickers are incredibly fast, so I’m pretty sure the InSight is working hard to keep up. Unfortunately, there is no video of the system, not even on the integrators web site, http://www.qcomptech.com/index.html. (They do have many other videos though; look here. )

I did notice though that the system is due to appear on the TV show, “How It’s Made” towards the end of 2010. Record it so you can watch the FlexPicker in slo-mo.

Adding 3D machine vision to robots

So, unable to exercise the brain on the golf course, you’ve been wondering how to make use of vision guided robotics? Well take a gander at this interesting blog post: “Robots Use 3D Vision to Improve System Costs and Product Quality” (posted Thursday Feb 11th on the “Everything about Instrumentation Engineering” blog.)

This provides an introduction to where and when to combine machine vision with robotics. Machine loading, bin picking and inspection are all covered. Just one word of warning though: the integrators and salespeople quoted all make it sound rather straightforward. It’s not.

VGR by webcam

Roborealm is a low cost machine vision software package popular among robotics hobbyists but I haven’t been aware of away one using it in a more industrial context.

Until now. On the blog “Profmason.com,” Professor Mason of Walnut, California has posted a tutorial, ”Controlling an Industrial Robotic Arm using Computer Vision” that describes how to use Roborealm and a webcam to control a Mitsubishi industrial robot.

This is a great little article for anyone interested either in figuring out what to do with Roborealm, or in learning more about vision guided robotics. If that’s you, use the links above to read more.

Great Vision-Guided-Robotics Video

I love seeing machine vision applications, and this has to be one of the most fun. The movie below shows ABB Flexpicker robots being used on a pancake packing line. It’s a four camera/four robot installation and the systems are running at a pretty good speed. Unfortunately though, and as always with YouTube videos, there’s a dearth of technical info.



We’re told that the cameras are GigE, but nothing about resolution or brand. Neither is there anything about the vision system. I’m guessing that it’s integrated into the robot controllers by ABB but I have to believe they’re buying the actual product from someone else, maybe Cognex.

For those who are interested in knowing who was actually responsible for all the engineering, the system was put together by RG Luma of the UK.

VGR Movie

“I love work. I can watch it all day,” or so said Oscar Wilde. He could have been talking about watching robots at work; something I find fascinating. When you see machines picking and placing at high speed it’s hard to avoid thinking there’s some intelligence buried inside the hardware.

This movie from Fanuc shows what I mean. Two parallel kinematic robots with vision guidance assemble and disassemble keyboards at high speed. It’s like watching ants at work. If you’re not fascinated you can’t be an engineer.

Picking your VGR technology

There are quite a few different ways of providing vision guidance to a robot; the challenge is in determining which is best for your application. There’s a good overview of the options in the September 2009 edition of Assembly, “Robotics: Guidance Goes 3D” (John Sprovieri, August 26, 2009.)

One thing to note though is that Vision Guided Robotics (VGR) is a broad tent covering many different types of task. Some of the easier jobs, (although that’s a relative term,) involve determining the exact position of a target so that the robot can orient itself for part pick up or placement. At the almost impossible end of the VGR spectrum lies the random bin picking job.

The Assembly article is, in my humble opinion, a little brief on the complexities of random bin picking. This has been the “Holy Grail” of VGR research for many years, and while I’ve no doubt that the practitioners – Braintech and Adenovation, for example – are inching closer to industrially robust systems, I’m still a skeptic. In a factory environment there are just so many environmental variables outside the control of the integrator.

I’m sure that in the next few years we’ll see a big increase in the adoption of VGR, but for now it’s a “bleeding edge” technology to be attempted only by the very bold.

Seriously high speed VGR

We’ve all seen robots playing pool, soccer and so on, but this takes some beating. See a three fingered robot bounce balls and play catch. I challenge you not to be impressed!

Starting out with VGR

For many engineers it’s the bin picking problem that first leads them in to Vision Guided Robotics, or VGR for short. “Wouldn’t it be nice,” they think to themselves, “to have a robot pick castings, (or forgings or molded parts,) from a bin and load them in to the next machine.”

Well yes, it would be nice, but it’s also quite difficult to do. If you’re just starting to think about such a VGR application let me refer you to this excellent primer, written by David Dechow of Aptura Machine Vision Solutions, and published in Vision & sensors, April 29, 2009.

Careful study of David’s paper may lead you to think that it might be cheaper maintain control of each part you produce, rather than jumbling them in a bin and then spending a lot of money to sort them out again. Don’t get me wrong; I love the technology, but a good engineer should always be looking for the easiest way to achieve his objective, and as David teaches us, VGR ain’t easy.

VGR Videos

I firmly believe the best way to demonstrate the capability of a vision system is through video. So I tip my hat to Braintech for placing four fascinating (although rather long,) movies on their web site. (You’ll find them on the right side of Braintech’s homepage.)

Braintech is a leader in vision guided robotics, and the movies demonstrate how they are pushing the envelope when it comes to bin picking. I’ve long been skeptical about just how reliable such a system can be in a factory environment, but I have to say that these Braintech videos go a long way to winning me over.

Memo to Braintech: much as I love the vids, please drop the cheesy musak. It reminds of some cheap movies my older brother used to bring home, if you know what I mean!

PoE and a VGR application

I’m not a big fan of alphabet soup, so let me start by explaining those two acronyms. PoE is of course Power over Ethernet, the latest flavor of the GigE camera interface, and VGR stands for Vision Guided Robotics. And with that introduction out of the way, let’s talk about why they go together like a horse and carriage.

The advantage of PoE over regular GigE is that only a single cable is needed. In this regard it makes a GigE interface much like 1394/Firewire. And that’s of great interest to anyone wanting to mount a camera on the end of a robot arm. Fewer cables mean less weight and less to break.

These points and more are explored in “A Vision of Robot Reliability,” authored by Max Falcone of Comau and published on ControlDesign.com. The article is also a big ‘plug’ for Comau’s new machine vision package, RecogniSense, which I’m not familiar with but might be worth checking out.

The bottom line is that marrying a robot with a vision system can help reduce downtime, and so more than justify the additional expense. Sounds good to me.