Hosted every year by the Danish Technological Institute (DTI), the RobotBrag event is the largest robotic exhibition in Denmark. Pressed by the public deadline, the institution had to come up with an "audience-wowing" bin-picking demo, find the best 3D camera, get familiar with a new vision system, and integrate it into a running system - in less than 3 weeks. Working against the clock was the biggest challenge and risk.
The automated bin-picking solution was composed of a Zivid One + 3D camera mounted on the arm of the robot, coupled with Halcon from MvTEC software and Zivid's Genicam interface. Thanks to the quality of the development team, the Zivid SDK, Halcon platform, and proven integration - plus the support from Zivid's Customer Engineering team, the solution was built in record time with flawless execution during the event.
The application was a great success for the DTI. With more than 3000 visitors, the bin-picking demo was particularly appreciated. As a positive side effect, the development team at the DTI acquired a set of new skills both for the Zivid products as well as for Halcon. It has demonstrated for them the robustness and capabilities of integration of the Zivid and the MvTEC product platform.
We only had three weeks to get familiar with both completely new hardware, as well as developing a new way to pick unknown objects from bins. The support from Zivid was impeccable and this was a very easy process that went very smoothly.” - Carsten Panch Isaksen, Senior Consultant
The future of technology was on display at the recent RobotBrag event. Hosted by the DTI and the Danish Robotic Association (DIRA), this exhibition is the largest robotic event in Denmark, attracting more than 3000 people along with 76 exhibitors this year.
This year, the DTI wanted to celebrate the 100th anniversary of the word “robot” with several innovative solutions including a bin-picking application. As a spokesperson of technological leadership in Scandinavia, the DTI wanted to set an example in bin-picking at RobotBrag. The demo would be an “order picking/order fulfillment” case where a robot would pick an object from three different bins and place them in an output bin. For this demo that would run for 2 full days, they needed the best 3D vision technology possible to succeed.
But there was a major issue: the event was only 3 weeks away.
The DTI only had 3 weeks to come up with a bin-picking demo, find a suitable 3D camera, and get familiar with a vision system they did not know. Working against the clock was the biggest challenge to overcome.
Furthermore, the elements to grasp were tricky. The robot had to pick both small and large objects in the bin, objects that could either be shiny metal, a transparent bag, or matte black. The variety of elements in shape, color, and material makes the task more difficult for the robot. Objects that are complex-shaped, hard to manipulate, entangled, and/or highly reflective machined or polished metals, are particularly hard to grab.
The robot cell also needed to be capable of delivering the physical reach capability, mechanical lifting capacity, and flexibility requirements demanded by the logistics application. It needed to be able to pick anything arriving on pallets and flow racks at a very high rate and on a continuous basis.
Finally, the robot had to traverse different bins and needed as much flexibility as possible. The vision system should not have knowledge about the contents in the bin, just know how many from each bin to pick. It is true for any application but in this situation, the choice of the vision system - and in particular the 3D camera – truly impacts the robot’s ability to successfully detect, pick, and place all types of parts.
The automated bin-picking solution for RobotBrag was composed of a LR Mate 200iD mini robot, coupled with our Zivid One + Color 3D camera. They used Zivid’s GeniCam interface for Halcon from MvTEC to segment the different parts in the bin. The DTI also implemented its own acquisition/segmenting solution based on its own framework. Finally, they used the calibration CLI from Zivid to perform the hand-eye alignment from 3D scanner to the robot flange.
Want to try it yourself? Download any component of our latest software here.
The limited amount of time available meant it was necessary for the customer to have a developer-friendly, adaptable 3D system that could be coupled with their framework in order to pick unknown objects easily. For Carsten Panch Isaksen, Senior Consultant, “We only had three weeks to get familiar with both completely new hardware, as well as developing a new way to pick unknown objects from bins. But this was a very easy process that went very smoothly”.
High-resolution 3D scanning with an acceptably low-point tolerance was also a key criterion for gripping the objects reliably and Zivid One Plus captures stunningly accurate and precise point clouds, in high-resolution and in color. Designed from the ground up to be developer-friendly, all Zivid 3D color cameras come with full customer support until you are completely happy with your machine vision deployment. Zivid provides extensive documentation, Knowledge Base, and GitHub code repositories. This was particularly appreciated by The DTI as they had so little time to get familiar with the product. Carsten Panch says “The support from Zivid was impeccable. Amy (CSU team) helped us the whole way answering questions quickly. There was a lot of useful documentation on the Knowledge Base to instruct on the acquisition variety. The visitors were absolutely delighted and blown away by the end result”.
To make a meaningful demonstration of the robot picking objects from different bins, the 3D scanner was also mounted on the arm of the robot. This made it possible to traverse different bins for maximum flexibility. Indeed, attaching the 3D camera directly onto the robot arm unlocks a much more flexible and effective solution since the 3D camera operates in its optimal space regardless of the working volume and robot reach. But until recently 3D vision sensors have just not been robust enough to withstand the mechanical demands and possibility of impact with on-arm robotics. With Zivid One Plus and Zivid Two, finally, these challenges are solved. Learn more about the benefits of 3D vision for on-arm robotics.
The system was constantly on for 2 full days, without a bug, only interrupted when there were no more products to pick. The application has been a success for the DTI: with more than 3000 visitors, the bin-picking demo was particularly appreciated. As a positive side effect, the development team at the DTI acquired a set of new skills both for the Zivid products as well as for Halcon. It has demonstrated for them the robustness and capabilities of integration of the Zivid and Halcon product platforms.
The Danish Technological Institute is a leading research and technology company. With more than a hundred years of existence and world-class equipment, they want to help their customers convert the newest knowledge and technology into value. They host the RobotBrag event at their HQ, attracting more than 3000 people every year.