Autonomous Machine Vision enables quality assurance for SMEs

Lacking the technical and financial resources of big enterprises, small to medium businesses (SMBs) might think that machine vision is not a viable option for them. Their perception is that automating quality assurance (QA) is complex and expensive. As a result, many SMBs rely on manual inspection and do not automate. Here Ofer Nir, CEO of Inspekto, explains why Autonomous Machine Vision represents a new era of industrial quality inspection, making automatic QA a viable and cost-effective option for SMBs, and not only for manufacturing giants.

SMBs are facing growing challenges. Smaller production lot sizes, frequent line changeovers, and increased competition are pressuring them to become increasingly efficient, agile, and productive. The growing difficulty to acquire skilled workers for QA positions is creating another challenge in this fast-paced environment.

This is why many small to medium manufacturers are moving towards automating production. Quality inspection must be automated as well to ensure that it is not a limiting factor in manufacturers’ ability to automate end-to-end, become more efficient and agile, and thus more competitive.

Current QA technologies and solutions are of limited use to small manufacturers. This is due to the fact that traditional machine vision projects are notorious for being complex, long, and expensive to set up and train, as well as rigid and tailored to the specific part in the production. On the other hand, SMBs are known to offer versatility and agility, which means constant diversification in the parts they produce or assemble. The result is that manual inspection is still the preferred QA method in many SMBs.

The shortcomings of manual inspection

The manual inspection comes with a series of pain points. The first is that in many cases, this method is based on sampling, meaning that for every inspected item, there are dozens of items that don’t undergo any checks at all.

The second is that manual inspection is a tedious, repetitive process, so inspectors could get tired or distracted and let their guard down, missing errors. Also, as with any manual task, different people might not have the same quality standards and so inspection may vary from one shift to another.

Finally, some items are simply too complex to be efficiently inspected by the human eye, meaning that hardly visible defects could easily go unnoticed through the QA process.

Autonomous Machine Vision

The system on a production line – the screen shows the area to be inspected

The solution to the shortcomings of a manual inspection lies in automating QA processes with a machine vision system that is cost-effective, easy to set up and train, and flexible enough to tackle a wide variety of QA challenges. An effective QA solution also needs to be generic or universal, meaning that it can inspect a wide range of parts and use cases and not be limited to the one product it was tailored to inspect, providing the flexibility required to switch from one part to another, or one location to another, quickly and easily.

The INSPEKTO S70 Gen.2, the only Autonomous Machine Vision (AMV) system currently on the market, offers all of these characteristics. Unlike traditional machine visions solutions, which are built ad hoc by an expert, the S70 Gen.2 is an off-the-shelf product that includes everything the user needs — camera, lighting equipment, PC, mounting system, software, and more. It is ready to use and can inspect a very wide range of products, in a variety of materials, produced through many different manufacturing and assembly processes. It can be deployed to inspect incoming goods at different production steps or at the end of the line and can be installed as a standalone station or integrated on the production line.

State of the art technology at your fingertips

Mimicking cognitive human vision and using the most advanced AI technology, the system can detect defects by being exposed to an average of only 20 OK (good) samples and does not require any NOK (defective) samples. AMV systems learn how a product is supposed to look and, based on their strong AI capabilities and millions of previous data entries, can flag any deviation from the memorized image. This leads to two main benefits. The first is that AMV systems can detect a wide range of defects, even unknown and unexpected ones, and not just a few ones they’ve been trained to recognize. The second is that the setup process is extremely simple because it requires only good items, which are abundant in any production line.

The whole setup process can be performed by the plant’s own personnel in as little as 45 minutes, making AMV systems the first choice for manufacturers that are looking to automate their QA processes. If required, additional help is available from the Inspekto team, on-site or remotely.

AMV systems eliminate the risks that come with traditional machine vision systems. Implementing an INSPEKTO S70 Gen.2 and creating the proof of concept (POC) will take users less time than they would normally need to just research and select the right components for any other machine vision project.

QA also for small batches

Historically, the cost and complexity of machine vision solutions meant that manufacturers needed high production volumes or very expensive products to justify the purchase of these systems. However, the cost-effectiveness of AMV systems makes them ideal for small production runs, as well as production lines that are frequently modified and reconfigured to produce new items.

Because of their ability to retrain quickly and easily, AMV systems offer manufacturers unlimited flexibility. To inspect a new product, the user simply needs to repeat the initial setup procedure and create a new profile. In 45 minutes, the system will be ready to inspect a completely different item. There is no need to adjust the lighting apparatus, change any of the sensors or adjust the camera settings. As the name suggests, Autonomous Machine Vision systems autonomously adjust the parameters of the electro-optical system to fit the new location, environment, and product.

The system can also inspect several different products in the same location, easily switching from one to the other in real-time. In this sense, the S70 Gen.2 effectively mimics the capabilities of a human inspector, who can inspect different products on the same line and quickly determine whether they are good or faulty, but with increased precision and reliability and with no need for breaks.

Data record

As opposed to human inspectors, AMV systems remember the details of each and every item they inspect, storing images in an archive that manufacturers can consult at any time.

This offers SMBs the invaluable chance to see exactly where and why defects happen and optimize their production processes accordingly. Over time, this can substantially increase production quality and have a positive impact on the bottom line of the business.

To stay competitive in the era of Industry 4.0, SMBs cannot overlook the importance of smart, automated quality inspection. For more information on how AMV can boost quality in your SMB, visit www.inspekto.com.