Image Processing Solutions for Real-Time Assembly Line Inspections

Although we are already shifting into Industry 4.0, a good number of manufacturing assembly lines are still relying on human eyes to find and analyze defective products down the conveyor line.

Although it is impossible to replicate what the human eye can see, at the same time, no human eye can replicate the speed, precision, and accuracy of systems like machine vision and robots. An assembly line empowered with these systems can perform real-time repetitive inspection routines without error and exhaustion.

Introducing a comprehensive Machine Vision system that can help maintain product quality down the assembly line. Deploy these vision solutions with the right embedded computing and cameras and you’ll likely save money and time in the long run.

Let’s discover more about the functionality and features of a Machine Vision systems used for Assembly Line Inspection”.

Maintaining Quality in the Assembly Line with a Vision System

One of the most popular areas where Machine Vision (MV) systems are being used today is in the manufacturing process, especially in the assembly line. The MV replaces the human eye, with a faster and more accurate automatic vision system that can acquire and process images of products passing by the line in real-time.

In the manufacturing floor, the vision system solution should be composed of the following components: a camera/sensor with the right illumination, an image processing platform with MV software, and an actuator such as a robot.

This system inspects the assembly line in real-time while also triggering action mechanisms like pass/fail that reject certain products with defects, irregularities, and other manufacturing flaws. Aside from a simple “yes or no” result, these systems can also be configured to provide sophisticated analysis and control of the entire assembly line, and some even fix or level defective products.

What product characteristics can be inspected with a vision system?

To ensure the quality of a product, a vision system can inspect the assembly lines and detect the following product characteristics:

  • Detecting presence. Simply detect whether an object or part of the product is there, for example, ring pulls on cans.
  • Determining whether a part is placed correctly, for example, package seals or cable wiring.
  • Reading labels, for example, printing mistakes, in mailboxes, letters, etc.
  • Finding irregularities or flaws, for example, scratches, contamination, production line mistakes, discoloration, etc. These slight irregularities are difficult to find because they are unexpected and random. Vision systems use something referred to as the “Golden template” to compare other items in real-time and find these irregularities.

If the system is placed in the right phase of the production line, it can help save a large number of resources. Instead of finding a flaw in a finished product at the end of the line, the vision system can detect the defective part mid-process, reject it, and notify about it. This way the product can be fixed and put back on the line.

The Vision System: How Does it Inspect Products in Assembly Lines?

An inspection solution varies from product to product and requirement. It can be challenging to specify the inspection process of a complex application. However, a simple inspection can be easily broken down into the following few steps.

Image Acquisition > Image Processing > Additional Actions.

  1. Image Acquisition. The vision system uses a single or group of IP cameras to automatically take images of every product down the assembly line. To help the system inspect the products as the human eye would, there must be the right surrounding lightning, camera lens, and shutter configuration, camera orientation, etc. Once the image is acquired, whether in 2D or 3D, it is sent to a local platform (or embedded computer) for further analysis and image processing.
  2. Image Processing. Machine vision algorithms can help detect specific edges and patterns within the obtained image to find specific objects, such as the ring pulls of cans. Once the object has been located within the image, an anchor point is created. This anchor point reduces the range for further image processing thus saving time and resources. From now on, different analyses can be applied to the isolated object, for example, positioning, orientation, identification through codes, comparison with templates, etc.
    But the vision system is only capable of doing all this at the same speed of the manufacturing assembly line with the help of embedded computers with image processing capabilities and machine-learned vision inspection.
  3. Results Evaluation and Action. Having immediate results at the same speed as the manufacturing process gives room for additional actions. The image processing results will determine the next action down the assembly line. If the results are “fail,” a physical system such as a robotic arm, located down the line will eventually either push the product away from the conveyor line, let it pass or if possible, fix the flaw.

Image Processing Solutions for Real-Time Automated Optical Inspections.

To maintain product quality control using machine vision, the images must be acquired, stored, and processed in real-time. On top of that, the system must also be capable of providing feedback or reporting to a central management station.

One of the key components of the vision system is a high-performing embedded computing system such as the LEC-2580 which can process images from IP camera data feeds and send requests to actuators, like robotics arms. This embedded computer is the brain that performs real-time image processing and machine vision algorithms.

The LEC-2580 acts as the vision inspection controller of the assembly line. It integrates various components to make this happen:

  • Intel Core high-performance processor.
  • Intel HD graphic for intensive graphics processing in machine vision applications.
  • I/O connectivity for multiple cameras and protocols for Programmable Logic Controllers (PLC).
  • Dual HDMI ports to monitor and HMI interfaces.
  • It can operate in a wide range of temperatures.

Final Words

The leading-edge technology in cameras, sensors, robotics, machine vision software, and real-time image processing platforms provide endless possibilities for inspection and quality control automation.

Together, these leading-edge components make a machine vision solution that is hard to ignore.

Machine vision systems go well beyond what the human eye can do. They can improve the accuracy, speed, and precision of the inspection process at assembly lines. These systems can help achieve zero defects down an assembly line and improve overall product quality.

But it all comes down to the brain that integrates imaging systems with actuators, the image processing platform. An embedded computer like the LEC-2580 is well-positioned for industrial image processing and quality control.

Image Processing Solutions for Real-Time Assembly Line Inspections was last modified: January 29th, 2020 by LEI Technology