Intelligent machine vision equipment for quality control

In the context of Industry 4.0, intelligent machine vision equipment plays a key role in quality control. These devices are designed to analyse images and videos automatically and accurately, enabling defect detection, product classification and verification of specific characteristics.

The implementation of the intelligent machine vision systems in quality control offers numerous advantages over traditional methods. Some of the most important benefits are:

1. Greater precision and objectivityIntelligent machine vision equipment can perform quality inspections accurately and consistently, eliminating human bias and reducing sorting or detection errors.
2. Increased speed and efficiencyThese systems can process images and video at high speed, enabling fast and efficient inspection of products. This is especially useful in high-speed production environments, where manual inspection would be unfeasible.
3. Early detection of defectsMachine vision equipment can identify defects or anomalies in real time, allowing corrective action to be taken in a timely manner and minimising the impact on production.
4. Cost savingsIntelligent machine vision systems can reduce the costs associated with manual inspection and defective products. By automatically detecting and sorting defective products, additional costs for rework or returns are avoided.
5. Data collection for further analysisIntelligent machine vision equipment generates large amounts of data during the inspection process. This data can be used to perform statistical analyses, identify trends and optimise production processes.

Intelligent machine vision equipment is an essential tool in the quality control in Industry 4.0. Its ability to process large volumes of visual data accurately and efficiently provides significant benefits in terms of accuracy, speed, efficiency and cost reduction. With its application, companies can improve the quality of their products and optimise their production processes, thus contributing to the advancement of Industry 4.0.

What is Intelligent Machine Vision?

To understand what Intelligent Machine Vision is, it is essential to understand that it is a set of components. One of the most important is artificial lighting, which must be controlled and specific to each vision application. In addition, it is necessary to study the sensitivity and the optimal sensor type for each case. The processing hardware must be high-performance, especially in industrial applications that require high speed. As for the software, it is possible to work with classical image processing or with neural networks. A hybrid solution can be optimal, as each type of processing is suitable for different applications. It is important to have a flexible product, which can be adapted to any hardware and software requirements. Therefore, artificial intelligence and neural network training are fundamental in Intelligent Computer Vision to adapt the processing software to each specific case.

How intelligent machine vision helps industrial companies

Intelligent machine vision can also help industrial companies in the early detection of faults or anomalies in production processes. By constantly monitoring and analysing captured images, vision systems can identify patterns or signals that indicate potential problems, allowing corrective action to be taken in a timely manner and avoiding costly production downtime.

In addition, intelligent machine vision can improve safety in the industrial environment. Object or person detection systems can alert to risky situations, such as the presence of foreign objects on the production line or unauthorised access to restricted areas. This helps to prevent accidents and maintain a safe working environment for workers.

Another relevant aspect is the manufacturing data analysisa. Intelligent machine vision allows large amounts of visual information to be collected and analysed in real time. This provides companies with a detailed view of their production, identifying patterns, trends and opportunities for improvement. The data collected can be used to optimise processes, predict potential problems, plan preventive maintenance and make informed decisions based on concrete data.

In addition, intelligent machine vision can facilitate the integration of systems in industry, as it can work in conjunction with other components of Industry 4.0, such as robotics and the Internet of Things. These systems can interact and communicate with each other, enabling more complete and efficient automation of industrial processes.

The industrial machine vision offers a wide range of benefits for industrial companies. From improved efficiency and quality, to early detection of faults, optimisation of costs and resources, and improved safety. By leveraging this technology, companies can achieve greater competitiveness, efficiency and profitability in today's industrial environment.

Intelligent machine vision technologies

In this diagram I show you what a typical machine vision system would look like. We are talking about a series of interconnected devices that are autonomous and have all the necessary components built into them. These devices are capable of capturing images in real time and processing them with algorithms that are embedded in them, which gives them autonomy and decision-making capacity independently of their integration in a more complex and global system.

These devices can be connected to an analytics platform that is located locally at the customer or in the cloud. Typically this communication is used for online analysis, image and data storage, visualisation and reporting. All industries' quantity departments also require this functionality.

An important issue is the ability of these systems to learn and improve over time. It is critical to connect our systems and be able to feed them with new cases, improve algorithms and implement them quickly and efficiently without disrupting training or the operability of production lines.

This is a very important requirement, as industries do not stop and many work three shifts (morning, afternoon and evening) on weekends. Production managers in factories don't like it if someone stops a line because we have to upgrade, for example, hardware.

Such applications can be used in almost any field, such as offer recognition, facial or character recognition, medical imaging, pattern recognition, quality control, defect detection in any type of part of any material (metal, plastic, wood), code reading and management of people or vehicle flow.

State of the art of intelligent machine vision

At the moment, there are four major trends in intelligent machine vision systems. Although there are more, in order not to go too deep, we will take these four as the most important:

• Embedded Systems.

• Deep Learning.

• Hyperspectral Imaging.

• 3D vision.

Embedded machine vision equipment

The miniaturisation of embedded machine vision equipment is revolutionising the way this technology is implemented in various fields of application. These compact and robust devices offer numerous advantages and possibilities in terms of versatility, portability and integration.

One of the main advantages of embedded machine vision equipment is its small size. Because they are so compact, they can be easily integrated into limited spaces or mobile devices, such as industrial machines, robots, autonomous vehicles, drones, medical devices, among others. This makes it possible to bring machine vision to places where it was not feasible before, opening up new opportunities and applications.

The portability of these devices is also a key factor. Being small in size and self-contained in terms of power and processing, they can be transported and used in different locations without difficulty. This facilitates deployment in changing environments or in applications that require mobility, such as field inspections, quality control on different production lines or remote monitoring.

In addition, embedded machine vision equipment is often highly rugged and resistant to harsh conditions. They are designed to withstand vibration, shock, extreme temperatures and harsh environments, making them ideal for industrial or outdoor applications where conditions are harsh.

The integration of these devices is another highlight. Having all the machine vision and data processing capabilities in a single compact device makes it easier to deploy and connect with other systems. They can be equipped with cameras, sensors, high-performance processors and specialised software, all integrated into a single device. This simplifies configuration, control and interaction with other system components, improving efficiency and facilitating integration with existing platforms.

Image processing equipment

As for the second important field, it is image processing. As I mentioned before, classical image processing is still being worked on as it is still more efficient in some applications than the use of neural networks. For example, dimensional and volumetric control, code reading and detection of parts on the production line. 3D vision is also used for robot guidance in the automotive sector and for quality control of parts.

However, when it comes to products with different parts, such as foodstuffs, wooden parts with different finishes, roughness and grain, plastics of various types, textiles and medical images, Deep Learning gains ground over classical image processing. In these cases, there is no particular pattern, so neural networks are more effective.

Deep Learning

In this field of deep learning, you will find yourself at the cutting edge of computer vision, as it is now rare to find an application that does not use some form of neural network to perform classification. In fact, in many of your applications, the implementation of neural networks is the norm. Even if it is a hybrid system that combines classical processing to prepare the classification to be performed using neural networks. That is why this paper is entitled Intelligent Computer Vision, because deep learning is playing an increasingly important role in this field.

Hyperspectral Imaging

Hyperspectral imaging is an advanced technology that captures detailed and accurate information about the chemical composition of objects and materials. Unlike conventional imaging that captures information in three channels (red, green and blue), hyperspectral imaging captures information in multiple bands of the electromagnetic spectrum, which provides a higher level of detail and allows the inference of characteristics and properties that are not visible to the naked eye.

Hyperspectral sensors operate over a wide range of wavelengths, from ultraviolet to near-infrared. These sensors capture the light reflected or emitted by an object in hundreds or even thousands of different spectral bands. Each band contains information about how the light interacts with the materials present in the scene, making it possible to identify specific characteristics, such as the presence of certain chemical elements, the concentration of compounds, the health of a plant or the detection of anomalies.

Hyperspectral imaging is commonly represented as a data cube, where each pixel in an image represents the spectral response at a specific wavelength. This data can be processed and analysed using advanced algorithms and techniques to extract valuable information and perform tasks such as change detection, material classification, chemical identification and defect or pest detection.

To process the large amount of data generated by a hyperspectral image, powerful processors and specialised algorithms are required. The amount of information contained in each image is significant, as multiple spectral bands are captured for each pixel, which implies intensive processing and the need to use specific data analysis techniques.

Hyperspectral imaging has applications in a wide range of industries and fields, such as agriculture, remote sensing, medicine, the food industry, geological exploration and environmental monitoring, among others. It provides detailed information on the chemical composition of objects and materials, facilitating decision-making, process optimisation and early detection of problems or anomalies.

3D Vision

In the field of machine vision, 3D technology is constantly evolving. There are several technologies available, some of which have been on the market for longer, such as binocular vision, which uses two cameras to simulate human vision. This technology allows us to see objects in three dimensions by processing and sharing information between the two cameras. There are also systems based on unstructured light, which project a laser pattern onto the object and analyse how it deforms and then make a three-dimensional reconstruction. Another method is laser triangulation, where the object is scanned with a laser and reconstructed in full in a few seconds.

Finally, there is the latest technology, known as «time of flight». This technology emits a wave and calculates the time it takes to receive the bounce of the wave, allowing a three-dimensional reconstruction of the part pixel by pixel. Each technology has its advantages and disadvantages, so it is important to consider what you need to choose the most suitable one. Factors such as distance, resolution, accuracy, image processing complexity, real-time compatibility, behaviour in different lighting conditions and cost should be carefully evaluated before making a decision.

Machine vision offers several options for working in three dimensions, and the important thing is to analyse the specific priorities and needs in order to select the most appropriate technology for each situation.

Application cases of intelligent machine vision equipment

In terms of successful application cases in different sectors, it is important to mention that virtually any field can benefit from machine vision. In fact, when you mention to someone that you are in this field, they will quickly think of an application in their work, environment or daily life.

For this reason, we can divide the application cases into four large blocks: Industry 4.0, the agri-food industry (both food processing and fresh food), flows of people or vehicular transport, and the manufacturing sector. We will focus mainly on the industrial sector, both in manufacturing and in the food sector.

The food sector is one of the strongest and most established at industrial level, which makes it very susceptible to take advantage of this type of technology. Each product is different and the quality control carried out by human factors is very weak, so the use of technologies such as artificial vision is very valuable and reinforces the quality of the products.

Machine Vision Equipment in the Manufacturing Industry

Industry 4.0 and machine vision systems offer a wide variety of applications in different industrial sectors. Some of the specific applications that can be carried out with these systems are:

• Inspection of surface defectsMachine vision systems can detect and classify surface defects on steel and aluminium profiles, such as scratches, dents, marks or irregularities. This helps to ensure product quality and prevent defective products from entering the market.
• 360 degree quality controlUsing advanced cameras and algorithms, vision systems can perform complete quality control on parts made of any material, such as plastic, metal, wood or glass. This includes detecting defects, measuring dimensions and features, and verifying the overall quality of parts.
• Robot guidance for parts handlingMachine vision systems can be used to guide robots in the handling and assembly of parts. Cameras can detect the position, orientation and characteristics of parts, allowing robots to perform tasks accurately and efficiently.
• Detection of residual stress in glass containersUsing polarised light technology, vision systems can detect internal residual stresses in glass containers, making it possible to identify defects that are not visible to the naked eye. This is especially important in applications where glass strength is crucial, such as in the pharmaceutical industry.
• Quality control in plastics packagingVision systems can perform inspections to verify the quality of plastic packaging, such as bottles, caps or containers. This includes detecting defects in shape, colour or labelling, ensuring that products meet quality and presentation standards.
• Inspection of capacitor blades and pneumatic tyresVision systems can detect defects in condenser sheets, such as cracks, bubbles or tears. In addition, they can inspect the quality and uniformity of tyres, identifying defects in the tread, sidewalls or the presence of foreign objects.
• Automated edition of panels and LEDs for the wood sectorVision systems can detect and recognise wood patterns on boards, enabling accurate automated editing. They can also inspect and classify LEDs in terms of quality and characteristics, facilitating the efficient production of lighting fixtures.
• Inspection of defects in steel structures and housesVision systems can detect defects or irregularities in steel structures, such as defective welds, corrosion or deformations. This helps to ensure the safety and quality of steel constructions and structures.
• Sorting and selection of productsVision systems can sort and select products based on their characteristics, such as size, shape, colour or labelling. This is especially useful in the food industry, where sorting of fruit, vegetables or packaged products according to quality or category can be carried out.
• Verification of assembly and mountingVision systems can verify that components have been assembled correctly and accurately in complex products. This is common in automotive manufacturing, where cameras are used to verify the assembly of parts on the production line.
• Inspection of weldsVision systems can inspect welds on metal products, identifying defects, cracks or non-uniformity. This is essential in the construction industry and the manufacture of metal structures.
• Quality control in printing and packagingVision systems can inspect print quality on labels, boxes and packaging, detecting printing errors, misalignment or visual defects. This ensures that products are properly presented and meet quality standards.
• Object detection and trackingVision systems can detect and track objects in real time, allowing them to be tracked and monitored. This is used in logistics and warehousing, where products or packages can be identified and tracked as they move through the supply chain.

Machine vision equipment in the food industry

In the food industry, various technologies and applications have been developed to improve the detection of both external and internal defects in fruits and other foods. These technologies use different channels, such as infrared or ultraviolet light, to identify and classify defects in products.

In the case of internal defects in fruit, such as bruises or damage not visible to the naked eye, scanning techniques are used to analyse the internal composition of the fruit. By detecting changes in the reflectance or absorption of light at different wavelengths, anomalies such as bruising, rot or internal damage can be identified without the need to open or damage the fruit.

These technologies are also used to sort products according to their state of ripeness. By analysing physical, chemical or biological characteristics, it is possible to determine the degree of ripeness of the fruit and to decide when it is the optimal time for harvesting or sale.

In addition, these techniques are useful in the handling of food products. For example, they are used for the detection of foreign bodies in food, such as stones or glass fragments, ensuring consumer safety. They are also applied in the evaluation of quality parameters, such as sugar content or brix degrees, which affect the taste and quality of food.

Another important area where these technologies are used is in crop management. In food processing plants, greenhouses and fields, applications are used to assist in harvesting, forecasting and estimating fruit growth. These tools help to optimise the resources used in harvesting, both in fruit trees and livestock crops, improving efficiency and reducing waste.

Food safety is also a crucial issue in the food industry. Specific applications are being developed to detect toxins in nuts, such as amygdalin in almonds, which is a serious problem affecting producers in several regions. These early detection technologies help prevent contamination and ensure product quality.

In addition to defects and food safety, it is essential to ensure quality control in packaging. Automatic label code reading systems are used and the expiry date and batch number are checked to avoid mislabelled or expired consignments reaching the final consumer.

Here are some examples of intelligent machine vision equipment for the food industry:

1. Infrared scannersare used to detect internal bruises on fruit without the need to open the fruit.
2. Ultraviolet scannershelp to identify the presence of mould or rot in foodstuffs.
3. Maturity classification systemssensors are used to determine the degree of ripeness of the fruit and to classify them according to their quality.
4. Gas sensorsare used to detect the presence of gases indicating the presence of decomposition or alteration of food.
5. Machine vision imaging technologyThe automatic detection and classification of external defects on fruit, such as spots, deformities or skin damage.

Machine vision equipment for people flow and management

Another large block of application sectors you can consider is the management and flow of people and vehicles. For example, in shopping centres and museums, people detection and counting, as well as identification through facial recognition, is important. In shopping centres, the ability to not only count or estimate capacity, but also to identify which profile of people is most focused on which type of product, is increasingly required. In museums, it is important to know which works attract which profile of people. For this, facial recognition is required, but not necessarily identification of individuals, as this may conflict with data protection issues. However, it is possible to obtain useful information from the person's profile, such as age, gender and race, for example.

It is also important to consider the identification and tracking of people to see where they are going, which is especially useful for security issues, such as in prohibited areas or if a backpack is left behind, for example. Other applications include the identification of points in shopping centres and public walkways, identification through number plate reading and tracking of containers to have traceability and to always know where the cargo is, which is especially important in port applications. Cameras can also be used for the guidance of container extraction cranes and to detect possible obstacles.

In addition to the detection of people, it is also possible to track and identify any type of vehicle, such as containers or transport vehicles. This is especially useful for companies that want to have traceability and always know where their cargo is. Cameras can also be used to read vehicle number plates, which is a very typical application.

Other applications include detecting the occupancy level of an enclosure, detecting the misuse of personal protective equipment by workers, measuring the volume of packages and reading labels in logistics warehouses, monitoring the activity and productivity of workers, and detecting any type of defect in fruit, both internal and external.

It is also important to consider the prevention of theft in shopping centres and to detect suspicious behaviour, such as leaving a backpack in a bus or underground car. In addition, panels in a logistics warehouse can be counted and tracked, contaminating products can be stopped, and fungi, bacteria or viruses can be detected on the surface. Although research is still ongoing, hyperspectral cameras may be useful for identifying some types of bacteria and fungi that are visible at certain frequencies in the infrared spectrum.

Suppliers of machine vision equipment for the manufacturing industry

At Technology for Industry we provide state-of-the-art machine vision solutions to optimise and improve the efficiency of production processes in your company. We can put you in touch with the leading machine vision technology providers that work best in terms of price, technical advice and lead times. They have a wide range of equipment and systems customised to meet the specific needs of your industry.

Their approach is based on understanding your specific challenges and requirements to provide you with solutions tailored to your needs. They work hand in hand with your production, operations and maintenance team to ensure successful implementation and seamless integration with your existing systems.

Machine vision technology uses advanced algorithms and high-precision sensors to perform fast and accurate inspections in real time. This enables early detection of defects, errors and anomalies on the production line, reducing costs associated with quality and improving overall productivity.

In recommending these suppliers, we stand behind their proven expertise and the quality of their solutions. They have demonstrated a consistent commitment to innovation and excellence, providing reliable, high-performance machine vision equipment.

In addition to offering quality equipment, these suppliers also provide excellent technical support and maintenance services. They are committed to working closely with your production, operations and maintenance team, ensuring successful implementation and seamless integration with your existing systems.

Relying on these machine vision providers will not only improve the efficiency and quality of your production processes, but also enable you to establish a competitive advantage in your industry. Their customised approach and ability to adapt to your specific requirements make them the ideal partner to drive your business success.

I invite you to explore the machine vision solutions that these providers have to offer. Do not hesitate to contact one of our advisors to discuss your needs and find out how they can help optimise your operations and improve your profitability.