Download PDF version Contact company

High demands on products as well as high time and cost pressure are decisive competitive factors across all industries and sectors. Whether in the food or automotive industry quality, safety and speed are today more than ever before factors that determine the success of a company. Zero-defect production is the goal. But how can it be guaranteed that only flawless products leave the production line? How can faulty quality decisions, which lead to high costs, be avoided? In order to test this reliably, a wide variety of methods are used in quality assurance.

A visual inspection with the human eye is possible, but it is often error-prone and expensive: the eye tires and working time is costly. A mechanical test, on the other hand, is usually accompanied by complex calibration, i.e. setting up and adjusting all parameters of both software and hardware in order to detect every error. In addition, product or material changes require recalibration.

Furthermore, with the classic, rule-based approach, a programmer or image processor must program rules specifically for the system to explain to the system how to detect the errors. This is complex and with a very high variance of errors often a hardly solvable Herculean task. All this can cost disproportionately much time and money. In order to make quality inspection as efficient, simple, reliable and cost-effective as possible, the German company sentin GmbH develops solutions that use deep learning and industrial cameras from IDS to enable fast and robust error detection. A sentin VISION system uses AI-based recognition software and can be trained using a few sample images. Together with a GigE Vision CMOS industrial camera from IDS and an evaluation unit, it can be easily embedded in existing processes.

Defect recognition software

The system is capable of segmenting objects, patterns and even defects. Even surfaces that are difficult to detect cannot stop the system. Classical applications can be found, for example, in the automotive industry (defect detection on metallic surfaces) or in the ceramics industry (defect detection by making dents visible on reflecting and mirroring surfaces), but also in the food industry (object and pattern recognition).

Depending on the application, the AI is trained to detect errors or anomalies. With the latter, the system learns to distinguish good from bad parts. If, for example, a surface structure is inspected, see metal part in the automotive industry or ceramic part, errors are detected by Artificial Intelligence as deviations from a comparison with reference images. By using anomaly detection and pre-trained models the system can detect defects based on just a few sample images of good parts.

The hardware setup required for the training and evaluation consists of an IDS industrial camera and appropriate lighting. The recognition models used are trained using reference images. For example, a system and AI model was configured for the error-prone inspection of fabric webs in the textile industry. A difficult task, as mistakes can be very subjective and very small. The system camera for optimum image material of textiles and web materials was selected together with IDS on the basis of specific customer requirements. A GigE Vision CMOS camera (GV-5880CP) was selected, which provides high-resolution data, triggered with precise timing, for accurate image evaluation.

IDS Vision CP camera

Slippage is the proportion of products that do not meet the standard but are overlooked

The system learns what constitutes a "good" fabric structure and knows already from a few shots of the fabric what a clean and flawless product looks like. For quality inspection, the image captured by the IDS Vision CP camera is then forwarded via GigE interface to an evaluation computer and processed with the recognition model. This computer can then reliably distinguish good/bad parts and highlight deviations. It gives an output signal when an error is found. In this way, slippage and pseudo rejects can be reduced quickly and easily.

Slippage is the proportion of products that do not meet the standard but are overlooked and therefore not sorted out, often leading to complaints. Pseudo rejects, on the other hand, are those products that meet the quality standard but are nevertheless incorrectly sorted out.

Both hardware and software of the system are flexible, for multiple or wider webs, additional cameras can easily be integrated into the setup. If necessary, the software also allows for re-training of the AI models. "Experience simply shows that a certain amount of night training is always necessary due to small individual circumstances. With pre-trained models from our portfolio, you need fewer reference images for individualisation and post training," explains Christian Els, CEO and co-founder of sentin.

GV-5880CP GigE Vision CMOS camera

Extremely accurate image acquisition and precise image evaluation are among the most important requirements for the camera used. Perfectly suitable is the GigE Vision CMOS camera GV-5880CP. The model has a 1/1.8" rolling shutter CMOS sensor Sony IMX178, which enables a very high resolution of 6.4 MP (3088 x 2076 px, aspect ratio 3:2). It delivers frame rates of up to 18 fps at full resolution and is therefore ideal for visualization tasks in quality control. The sensor from the Sony STARVIS series features BSI technology (backside illumination) and is one of the most light-sensitive sensors with a low dark current close to the SCMOS range (Scientific CMOS). It ensures impressive results even under very low light conditions.

"In addition to resolution and frame rate, the interface and the price were decisive factors in the decision for the camera. The direct exchange with the IDS development department has helped us to reduce the time needed for camera integration," says Arkadius Gombos, Technical Manager at sentin. The integration into the sentin VISION system is done via GenTL and a Python interface.

Advantages of AI over human inspection or applications

The human brain is beaten by artificial intelligence in terms of consistency and reliability

Automated, image-based quality control with Artificial Intelligence offers many advantages over human visual inspection or conventional machine vision applications. "In AI-based image interpretation, the aim is to create images on which humans can see the error, because then the AI model can do it too," concludes Christian Els. The system learns to recognize the requirements of the product similar to a human being. But the human brain is beaten at any time by an artificial intelligence in terms of consistency and reliability.

Even if the brain is capable of remarkable peak performance, an AI can recognize much more complex error patterns. The human eye, on the other hand, cannot stand up to any camera in terms of fatigue and vision. In combination with deep-learning recognition software, the image processing system therefore enables particularly fast and accurate inspection. Depending on the application, image acquisition and evaluation can take place in just a few milliseconds.

Deep learning

The system can also be applied to other areas such as surface testing. Similar applications are e.g. the testing of matte metal/coatings surfaces (automotive interior), natural materials (stone, wood) or technical textiles such as leather. Scratches, cracks and other defects on consumer goods can thus be detected and the respective products sorted out. Exclude quality defects and produce only good stuff, an indispensable process within the framework of quality assurance.

IDS cameras in combination with the deep learning supported software of sentin GmbH significantly optimize the detection of defects and objects in quality control. This allows the personnel and time expenditure for complaints and rework, as well as pseudo rejects, to be significantly reduced in a wide range of industries and areas.

Share with LinkedIn Share with Twitter Share with Facebook Share with What's App Share with Facebook
Download PDF version Download PDF version

In case you missed it

What You Need To Know About Open Source Intelligence (OSINT) For Emergency Preparedness?
What You Need To Know About Open Source Intelligence (OSINT) For Emergency Preparedness?

Have you ever stopped to consider the volume of new data created daily on social media? It’s staggering. Take Twitter, for instance. Approximately 500 million tweets are published every day, adding up to more than 200 billion posts per year. On Facebook, users upload an additional 350 million photos per day, and on YouTube, nearly 720,000 hours of new video content is added every 24 hours. While this overwhelming volume of information may be of no concern to your average social media user posting updates to keep up with family and friends, it’s of particular interest to corporate security and safety professionals who are increasingly using it to monitor current events and detect potential risks around their people and locations—all in real-time. Meet the fast-paced and oft-confusing world of open-source intelligence (OSINT). What is Open Source Intelligence (OSINT)? The U.S. Department of State defines OSINT as, “intelligence that is produced from publicly available information and is collected, exploited, and disseminated promptly to an appropriate audience to address a specific intelligence requirement.” The concept of monitoring and leveraging publicly available information sources for intelligence purposes dates back to the 1930s. The British Broadcast Corporation (BBC) was approached by the British government and asked to develop a new service that would capture and analyze print journalism from around the world. Monitoring and identifying potential threats Originally named the “Digest of Foreign Broadcast, the service (later renamed BBC Monitoring which still exists today) captured and analyzed nearly 1.25 million broadcast words every day to help British intelligence officials keep tabs on conversations taking place abroad and what foreign governments were saying to their constituents. OSINT encompasses any publicly accessible information that can be used to monitor and identify potential threats Today, OSINT broadly encompasses any publicly accessible information that can be used to monitor and identify potential threats and/or relevant events with the potential to impact safety or business operations. The potential of OSINT data is extraordinary. Not only can it enable security and safety teams to quickly identify pertinent information that may pose a material risk to their business or people, but it can also be captured by anyone with the right set of tools and training. OSINT for cybersecurity and physical threat detection Whether it be a significant weather event, supply chain disruptions, or a world health crisis few saw coming, the threats facing organizations continue to increase in size and scale. Luckily, OSINT has been able to accelerate how organizations detect, validate, and respond to these threats, and it has proved invaluable in reducing risk and informing decision-making – especially during emergencies. OSINT is typically shared in real-time, so once a situation is reported, security teams can then work on verifying critical details such as the location or time an incident occurred or provide the most up-to-date information about rapidly developing events on the ground. They can then continue to monitor online chatter about the crisis, increasing their situational awareness and speeding up their incident response times. OSINT applications OSINT can help detect when sensitive company information may have been accessed by hackers  Severe weather offers a good example of OSINT in action. Say an organization is located in the Great Plains. They could use OSINT from sources like the National Weather Service or National Oceanic and Atmospheric Administration (NOAA) to initiate emergency communications to employees about tornado warnings, high winds, or other dangerous conditions as they are reported. Another common use case for OSINT involves data breaches and cyber-attacks. OSINT can help detect when sensitive company information may have been accessed by hackers by monitoring dark web messaging boards and forums. In 2019, T-Cellphone suffered a data breach that affected more than a million customers, but it was able to quickly alert affected users after finding their personal data online. OSINT is a well-established field with countless applications. Unfortunately, in an ever-changing digital world, it’s not always enough to help organizations weather a crisis. Why OSINT alone isn’t enough? One of the core challenges with leveraging OSINT data, especially social media intelligence (SOCMINT), is that much of it is unstructured and spread across many disparate sources, making it difficult to sort through, manage, and organize. Consider the social media statistics above. Assuming a business wanted to monitor all conversations on Twitter to ensure all relevant information was captured, it would need to both capture and analyze 500 million individual posts every day. Assuming a trained analyst spent just three seconds analyzing each post, that would amount to 1.5 billion seconds of labor—equivalent to 416,666 hours—just to keep pace. While technology and filters can greatly reduce the burden and help organizations narrow the scope of their analysis, it’s easy to see how quickly human capital constraints can limit the utility of OSINT data—even for the largest companies. Challenges with OSINT OSINT data collection includes both passive and active techniques, each requiring a different level of effort and skill Additionally, collecting OSINT data is time-consuming and resource-intensive. Making sense of it remains a highly specialized skill set requiring years of training. In an emergency where every second count, the time required to sift through copious amounts of information takes far longer than the time in which an organization must take meaningful action to alter the outcome. Compounding the issue, OSINT data is noisy and difficult to filter. Even trained analysts find the need to constantly monitor, search, and filter voluminous troves of unstructured data tedious. Artificial intelligence and machine learning have helped weed through some of this data faster, but for organizations with multiple locations tasked with monitoring hundreds or thousands of employees, it’s still a challenging task. Adding to the complexity, collecting OSINT data isn’t easy. OSINT data collection includes both passive and active techniques, each requiring a different level of effort and skill. Passive vs Active OSINT Passive OSINT is typically anonymous and meant to avoid drawing attention to the person requesting the information. Scrolling user posts on public social media profiles is a good example of passive OSINT. Active OSINT refers to information proactively sought out, but it often requires a more purposeful effort to retrieve it. That may mean specific login details are needed to access a website where information is stored. Lastly, unverified OSINT data can’t always be trusted. Analysts often encounter false positives or fake reports, which not only take time to confirm accuracy, but if they act on misinformation, the result could be damage to their organization’s reputation or worse. So, how can companies take advantage of it without staffing an army of analysts or creating operational headaches? A new path for OSINT Organisations can leverage the benefits of OSINT to improve situational awareness and aid decision-making Fortunately, organizations can leverage the benefits of OSINT to improve situational awareness and aid decision-making without hiring a dedicated team of analysts to comb through the data. By combining OSINT data with third-party threat intelligence solutions, organizations can get a cleaner, more actionable view of what’s happening in the world. Threat intelligence solutions not only offer speed by monitoring for only the most relevant events 24/7/365, but they also offer more comprehensive coverage of a wide range of threat types. What’s more, the data is often verified and married with location intelligence to help organizations better understand if, how, and to what extent each threat poses a risk to their people, facilities, and assets. In a world with a never-ending stream of information available, learning how to parse and interpret it becomes all the more important. OSINT is a necessary piece to any organization’s threat intelligence and monitoring system, but it can’t be the only solution. Paired with external threat intelligence tools, OSINT can help reduce risk and keep employees safe during emergencies and critical events.

Baltimore Is The Latest U.S. City To Target Facial Recognition Technology
Baltimore Is The Latest U.S. City To Target Facial Recognition Technology

The city of Baltimore has banned the use of facial recognition systems by residents, businesses and the city government (except for police). The criminalization in a major U.S. city of an important emerging technology in the physical security industry is an extreme example of the continuing backlash against facial recognition throughout the United States. Facial recognition technology ban Several localities – from Portland, Oregon, to San Francisco, from Oakland, California, to Boston – have moved to limit use of the technology, and privacy groups have even proposed a national moratorium on use of facial recognition. The physical security industry, led by the Security Industry Association (SIA), vigorously opposed the ban in Baltimore, urging a measured approach and ‘more rational policymaking’ that preserve the technology’s value while managing any privacy or other concerns. Physical security industry opposes ban In such cases, it is local businesses and residents who stand to lose the most" “Unfortunately, an outright ban on facial recognition continues a distressing pattern in which the clear value of this technology is ignored,” said SIA’s Chief Executive Officer (CEO) Don Erickson, adding “In such cases, it is local businesses and residents who stand to lose the most.” At the national level, a letter to US President Biden from the U.S. Chamber of Commerce Coalition asserts the need for a national dialog over the appropriate use of facial recognition technology and expresses concern about ‘a blanket moratorium on federal government use and procurement of the technology’. (The coalition includes Security Industry Association (SIA) and other industry groups.) The negativity comes at a peak moment for facial recognition and other biometric technologies, which saw an increase of interest for a variety of public and business applications, during the COVID-19 pandemic’s prioritization to improve public health hygiene and to promote ‘contactless’ technologies. Prohibition on banks, retailers and online sellers The ordinance in Baltimore prohibits banks from using facial recognition to enhance consumer security in financial transactions. It prevents retailers from accelerating checkout lines with contactless payment and prohibits remote online identity document verification, which is needed by online sellers or gig economy workers, according to the Security Industry Association (SIA). At a human level, SIA points out that the prohibition of facial recognition undermines the use of customized accessibility tools for disabled persons, including those suffering with blindness, memory loss or prosopagnosia (face blindness). Ban out of line with current state of facial recognition Addressing the Baltimore prohibition, the Information Technology and Innovation Foundation reacted to the measure as ‘shockingly out of line with the current state of facial recognition technology and its growing adoption in many sectors of the economy’. Before Baltimore’s decision to target facial recognition, Portland, Oregon, had perhaps the strictest ban, prohibiting city government agencies and private businesses from using the technology on the city’s grounds. San Francisco was the first U.S. city to ban the technology, with Boston, Oakland; Cambridge, Massachusetts; and Berkeley, California, among others, following suit. Police and federal units can use biometrics Unlike other bans, the Baltimore moratorium does not apply to police uses Unlike other bans, the Baltimore moratorium does not apply to police uses, but targets private uses of the technology. It also includes a one-year ‘sunset’ clause that requires city council approval for an extension. The measure carves out an exemption for use of biometrics in access control systems. However, violations of the measure are punishable by 12 months in jail. The law also establishes a task force to evaluate the cost and effectiveness of surveillance tools. Transparency in public sector use of facial recognition Currently, the state of Maryland controls the Baltimore Police Department, so the city council does not have authority to ban police use of facial recognition, which has been a human rights concern driving the bans in other jurisdictions. A measure to return local control of police to the city could pass before the year lapses. SIA advocates transparency in public-sector applications of facial recognition in identity verification, security and law enforcement investigative applications. SIA’s CEO, Don Erickson stated, “As public sector uses are more likely to be part of processes with consequential outcomes, it is especially important for transparency and sound policies to accompany government applications.”

What Are The Security Challenges Of Protecting Critical Infrastructure?
What Are The Security Challenges Of Protecting Critical Infrastructure?

Many of us take critical infrastructure for granted in our everyday lives. We turn on a tap, flip a switch, push a button, and water, light, and heat are all readily available. But it is important to remember that computerized systems manage critical infrastructure facilities, making them vulnerable to cyber-attacks. The recent ransomware attack on the Colonial Pipeline is an example of the new types of threats. In addition, any number of physical attacks is also possibilities. We asked this week’s Expert Panel Roundtable: What are the security challenges of protecting critical infrastructure?