Vision Systems

Navigating Four Vision System Challenges in Bin-Picking Systems

On the surface, the process of a robot picking up and removing a target from unorganized or randomly arranged materials may seem “simple.” However, in practice, detecting, picking, and placing items from a bin requires vision systems to be programmed with all potential scenarios in mind.

A machine vision system is required to guide robots in their pick-and-place tasks when parts are not always in repeatable positions. Especially with traditional bin-picking solutions, the target’s image quality, limited rule-based algorithm, and many other variables could cause errors, which, when accumulated, lead to high uncertainty and missed picks.

As automation has advanced, however, modern vision systems from KEYENCE resolve many of the longstanding challenges posed by traditional bin-picking methods.

Understanding Bin-Picking Challenges

Automated random bin picking is the process in which a robot without a human actor identifies, picks, and removes a target from other objects in an environment or a “bin.” For example, the automotive industry uses 3D bin picking to pick randomized heavy parts that would tire operators down, and the distribution industry uses bin-picking to assist with order fulfillment.

Here are four common bin-picking challenges:

Precision and Trueness

Some vision systems help robots recognize targets but struggle with true object positions.

Speed

Limited processing power, poor lighting, and data transfer delays can impact speed in traditional bin-picking solutions.

Reflective Material Handling

Metallic and shiny objects challenge traditional bin-picking methods, causing robots to pick non-existent items or from the wrong place as glare shields the vision system from truly seeing the part.

System Stability

Vision systems in harsh conditions may struggle to perform, as 3D functions are impacted by the environment.

Ask an Expert

Addressing Challenges with Bin-Picking Vision Systems

Rather than serving as a useful complement to repetitive task handling, a vision system could become an impediment.

Fortunately, KEYENCE provides machine vision systems that address the above bin-picking challenges that are commonplace in older vision systems. Our solutions help engineers and developers have a seamless 3D robotic bin-picking experience.

High-Resolution Cameras

For example, the KEYENCE RB Series is a 3D vision system that provides unmatched detection performance. Four high-resolution cameras and a powerful projection laser light let the system capture high-quality images from all angles, eliminating blind spots and related problems. The imaging unit can capture up to 136 images in a fraction of a second, from which the 3D module can make calculations to identify parts and the easiest pick. This enables the system to generate accurate 3D data of the target object, making it easier for the robot arm to pick and move it without any errors or collisions.

Automatic Path Planning

Automatic path planning design prevents ghost points. The CV-X Series automatically calculates the path for the robot so the arm doesn't interfere with obstacles or interact with other things. All simulations are conducted on a PC, eliminating the need for manual processes. With the built-in Picking Simulator, users can test out different grippers or cell designs and simulate the picking results before implementing them in the actual work cell. This allows for thorough planning and optimization of the entire automation process, resulting in the elimination of rework waste due to poor designs.

1

Automatically calculates paths that avoid obstacles
2

Automatically calculates paths where the arm doesn’t interfere with the container

Simple Setup

Another key feature of this system is its simple setup process, which includes automatic robot-camera calibration and the option to upload CAD data for an even faster setup. This not only saves time but also eliminates the need for highly skilled operators, making it accessible to a wide range of manufacturing facilities.

1

Projector
2

4 cameras

Key features of KEYENCE 3D vision guided-robotics solutions include:

Four high-resolution cameras and a complementary projector
High-speed processor
High-precision CMOS sensor
Optimal image-capturing angle
Newly developed 3D scanning function
Automatic path planning function
Picking simulator
HDR image function
Multiple reflections suppress function

With the increasing demand for automation in manufacturing processes, it has become crucial to have efficient and reliable systems that can accurately detect and handle objects. The 3D vision-guided robotics system from KEYENCE offers just that: a powerful tool for automating tasks such as assembly, de-palletizing, and machine tending.

Price

FAQs About Bin-Picking

How Do Vision System Challenges Impact the Effectiveness of Bin-Picking Solutions?

Vision system challenges like precision, trueness, system stability, and reflective material handling could lead to errors in bin-picking tasks. If, for example, a robot can recognize a target but not the target’s true location, this leads to collision with surrounding objects and overall failure of the system. Excessive downtime of a vision system would also mean pick-and-place tasks not getting done as and when due, which may impact overall production throughput.

What Are the Limitations of Traditional Bin-Picking Methods, and How Can KEYENCE Vision Systems Address Them?

Traditional bin-picking methods present several limitations. Many of these solutions require manual calibration, a process prone to human error and often time-consuming. Furthermore, making adjustments can also be challenging.
KEYENCE vision systems solve these issues by providing fully automated robot-camera calibration and easy programming with CAD data upload capabilities. This not only reduces the potential for error but also speeds up the setup process significantly.

How Do Bin-Picking Vision Systems Contribute to Reducing Errors and Optimizing Production Processes in Various Industries?

Modern, robust bin-picking vision systems offer automatic calibration and readjustment, resulting in fewer errors and better system reliability. The automatic path planning function in modern vision systems also helps robots and devices navigate obstacles and precisely place objects without manual programming or human input.

What Are the Main Obstacles Encountered in Bin-Picking Operations, Particularly in Complex Manufacturing Environments?

Problems with object recognition, precision of 3D points, speed of acquisition, reflective object handling, and system stability are the major obstacles faced in complex manufacturing sites. Modern, high-performance machine vision systems provide superior solutions to these problems.

Ready to address your industrial bin-picking problems? KEYENCE 3D vision-guided robotics provide unparalleled bin-picking solutions for even the most challenging tasks and complex manufacturing needs. Our products can be deployed afresh or integrated into existing warehouse logistics, order fulfillment, assembly lines, and many more.

Contact KEYENCE today to learn more.