Vision & Imaging Blog
Picking Objects with Vision Guided Robots Systems
Among the most common tasks in the industrial environment is picking objects. Objects often must be picked and placed to move them from one stage of industrial processing to another. Robots have proven effective in pick and place environments where all objects are of a similar size and their orientation can be reliably predicted.
As vision-guided robotics continues to mature, however, engineers explore new applications.
Most of these center on picking and placement of semi-structured objects. When items are not uniform in size, appearance, orientation, or relative location, how can robots be expected to interact with them? This major challenge is made up of many smaller issues.
Vision-Guided Robotics as a Solution to Ad-Hoc Picking and Placement
To pick semi-structured objects, the robot must evaluate them on multiple levels:
- Recognizing the right object in a collection of many irregular ones;
- Calculating the most efficient order in which to pick up the objects;
- Planning the right way to grip them, ideally without damaging them;
- Determining how to lift, move, and place them in the environment.
Until recently, subtle differences between similar objects represented a virtually impossible barrier to unstructured pick and place systems, even using cutting-edge machine vision. A recent project by Cambridge Consultants has moved the needle on this aspiration.
Through the talents of more than 500 engineers and other specialists, Cambridge Consultants developed a system that autonomously identifies different fruit and executes a complete pick and place operation by comparing colors.
Certified System Integrator Program
Set Yourself at the Forefront of the Global Vision Market
.jpg)
Vision system integrators certified by A3 are acknowledged globally throughout the industry as an elite group of accomplished, highly skilled and trusted professionals. You’ll be able to leverage your certification to enhance your competitiveness and expand your opportunities.
Although the system is intended only to demonstrate the concept, it has overcome many of the main technical obstacles. It uses infrared pattern recognition to locate the objects, computes shape, color, size, and depth information, analyzes both color and depth, then develops a 3D positional reference for each object – allowing it to plan gripper movement and trajectory.
Unstructured Pick and Place Could Soon Come to a Farm Near You
It’s no coincidence that fruit has been an inspiration for unstructured pick and place research. In addition to being an example of variable natural forms that artists have looked to for years, fruit represents a major area where agricultural machine vision is poised to grow.
Vision systems have been deployed to automate fresh fruit picking as far back as 2007. Still, there remained concerns about the accuracy and dexterity of such systems. Many fresh fruits must be picked carefully to maintain suitability for sale.
Now, both machine vision and robotic motion have advanced considerably. With guidance from sophisticated vision systems, robots could simulate human agricultural techniques more easily than ever – using, for example, the soft robot hand.
Thanks to these breakthroughs, all the pieces are in place to automate harvesting in new ways in the coming decade.
Recent Posts