Machine Vision Guide For Automotive Process Automation

Machine Vision Guide For Automotive Process Automation

RAGHAVA KASHYAPA is Managing Director of Qualitas Technologies

INTRODUCTION

The automotive industry is perhaps the most advanced in the adoption of precision automation technology. Many reasons contribute to this; the primary one being the high quality standards that are demanded in the production cycle. Automotive OEMs hold their Tier I suppliers to the highest levels of quality standards, levying high penalties for quality rejections and delayed supplies, which in turn is passed down to Tier II suppliers and so on.

With human errors being one of the primary factors impacting quality, it's no wonder that the automotive industry is at the forefront in adapting innovative quality control automation processes like poka-yoke in their processes. Machine vision is one such technology that has made significant forays into the industry, especially in the Indian automotive segment. The key advantage of using machine vision is in the fact that it uses a "non-contact" based inspection technique, thus making it easy to integrate into your production process without much modification to your assembly lines.

The primary use of machine vision systems in the factory is either to improve quality or to automate production. When used for quality inspection, the primary purpose of machine vision is to automate complex mundane tasks that can be performed at high speeds, with accuracy and consistency. Additionally, machine vision can be used to automate production processes like guiding a robot to perform a pick or a place operation by identifying the position of a target object.

KEY ELEMENTS OF A MACHINE VISION SYSTEM

A common mistake people make is to over-emphasise the role of the camera in a machine vision system. Although a good camera resolution is important, it's often not the most critical part in ensuring system accuracy. Lighting and good optics are often more critical. To understand the key elements is important to ensure system performance. A good system integrator should study the problem and requirements, often conducting lab and field trials to ensure that the design choices being made across all elements are the optimum ones.

MACHINE VISION APPLICATIONS

The first step in evaluating a solution is in the understanding of the problem itself. Make a list of the defects that you would like to identify, taking time to list out objectively what constitutes an unacceptable and what constitutes an acceptable defect. For example, if the objective is to identify scratch marks in a part, be clear to indicate specific dimensional details of the scratch, especially if some scratches are acceptable and not to be classified as defects. If there are multiple variants of the same part, be sure to compile a list of all variants of the part; if it's a forged part, list all variants and their sizes and drawings.

Functionally, machine vision applications can be classified into a few application groups. It's important to understand the kind of application your requirement falls under so as to decide which kind of system design approach you need to invest in. Often, there will be a need to have one (or even more) functional requirements that your applications demands. Listed below are the primary application categories:

Verification: Here the objective is to check whether a part is present or absent. An example is to find whether a circlip or bearing is present on an engine assembly. The vision system simply identifies the part by virtue of a "golden image" that has been saved, which it uses to compare it with the real time images from the camera.

Measurement applications are used where the objective is to measure certain features within the captured image. It could be dimensional measurement of machined parts like diameters, length, pitch diameter, etc. Measurement can be made either with 2D or 3D imaging technologies.

Flaw detection applications detect abnormalities such as surface defects, dents and scratches on a product surface. Flaw detection applications need to be carefully objectified in order to ensure "acceptable" flaws can be distinguished from unacceptable flaws.

Print Defect Detection: Identification of printing anomalies like incorrect colour shades or where parts of the print is missing or blemished is the objective of print defect identification. Here, a golden or master image is trained to the system in order to identify any deviations from this master. Examples are printed matter for automotive dashboards.

Identification using machine vision involves identifying a part or product in order to track this part across the manufacturing or logistical process or to verify that the right part is being produced. Identification can be done either by reading characters (OCR) or barcodes.

Locating objects is a common use of machine vision for applications such as robotic guidance. Here, the machine vision system's objective is to locate the coordinates/ position of an object of interest. This information can be used to, say, pick-up the object or perform any other process that is dependent on this location. Another example is to locate the weld seam of an automotive wheel rim.

Counting (DR) is the use of machine vision to count objects of interest (say, counting number of piston rings on a stack).

14-15 guestCommentary atrAug16 9

INTEGRATION CONSIDERATION

After the problem has been defined, it's important to understand the practical aspects of how to integrate this into your production process.

Online or Offline: When conceptualising the use of machine vision in your production line, the primary choice one has to make is whether to integrate the system in-situ or in-line with the production process or to have a standalone system. Given that the primary advantage of machine vision is mistake proofing, integration in the production process would be most beneficial, provided it can be done. However, if not planned or designed well, machine vision systems can also cause a lot of disruption in the process like false rejections, production stoppages or worse, accept defective parts.

Control Automation: Keep in mind that a vision system acts as a "sensor" that identifies a defect or some element in the part, which in turn drives some kind of automation. It would help define what downstream automations are required like line-stoppage, part ejection, buzzer or even SMS notifications or alerts. After all, you want to gain maximum use of this investment to ensure your automation requirements are fully met. Arriving at clarity in this helps define the scope of work clearly and also allows you to plan for integration work that will be required, reducing surprises during system integration.

CONCLUSION

Machine vision is a hot area and will make a huge impact in automotive process automation in the coming times. As a buyer or a user, it's important to be informed of the technology and not fall into marketing traps that often lead to undesirable results. As much as the technology offers promise, it's also important to note that it has limitations, and understanding and setting the right expectations within the organisation is critical in ensuring success. Choosing the right system integrator will also make sure that these important aspects are covered and in turn surprises are avoided later on. As with all technology, if it's applied with thought and preparation, it can lead to marvellous results. With intense competition, machine vision is one of the ways you can stay ahead of the curve by exploiting the advantages it has to offer.