Vision Sensors

Principles of vision sensors

The image captured by the camera in a vision sensor passes through the lens and is converted to an electrical signal by the light-receiving element. The brightness and shape of the target are then determined according to the light/darkness and intensity information separated into the number of pixels of the light-receiving element.
The light-receiving element most commonly used in vision sensors is a CMOS, a semiconductor that acts as an “electronic eye” to enable high-speed transmission by amplifying and directly detecting the signals from each pixel. Because each pixel has its own amplifier, electrical noise generated when reading the electrical signals can be suppressed. Compared to CCDs—image sensors with high-voltage analog circuits—CMOS consume less power thanks to their digital circuitry, and smearing and blooming are not generally a problem.
These advantages make vision sensors a popular choice in various manufacturing sites thanks to their compact size and stable, high-speed detection performance.

Types of vision sensors

Vision sensors are available in two types: monochrome or color. Both types convert brightness captured by the lens into electrical signals and use the light/darkness and intensity information to determine the contour, brightness, and shape of a target. However, the structure of the light-receiving element and the image processing principles between the two types are different, and the characteristics of each differentiation method also differ.

Monochrome vision sensors:
Monochrome vision sensors do not include a Bayer filter used to extract red, green, and blue (RGB) colors. This means the images captured by a monochrome vision sensor are clear black-and-white images with no grating caused by the Bayer filter.
Such monochrome sensors offer high-speed reading and stable detection for applications where color information is not required, such as when reading barcodes, measuring external dimensions, matching patterns, or counting targets. If, however, color images are sometimes needed for differentiation, a color vision sensor is recommended, as a Bayer filter cannot be attached to a monochrome sensor.

Color vision sensors:
Color vision sensors are equipped with a Bayer filter that separates the light captured through the lens into RGB color. The Bayer filter is attached in front of the light-receiving element and is connected to each pixel. As light reflected by the target passes through the Bayer filter, the individual pixels on the light-receiving element acquire either R, G, or B color information. After the acquired color information is converted into electrical signals, the signals are used to create a single image through a process called Bayer conversion.
This enables color vision sensors to accurately determine not only external dimensions, shapes, and quantity but also minor differences between colors and intensity.

Contact-type displacement sensors and laser displacement sensors capture information from a specific point on the target, so as many sensors as detection points are required. This results in significant effort installing a large number of sensors and positioning them to the desired locations. In addition, detection is only possible while the target is stopped.
Vision sensors, on the other hand, determine the target shape based on the image captured by the camera. The detection location is then specified using the image, eliminating the need for positioning equipment or for stopping the target. Moreover, the ability to perform detection of multiple points within the camera’s field of view at the same time means there is no need for a large number of sensors as with contact-type displacement sensors and laser displacement sensors. Being able to keep the number of necessary sensors to a minimum significantly reduces the work required for product changeovers.

The most common use for color sensors is color differentiation. However, small color differences and color irregularities in the detection area frequently cause false detections with some color sensors. Glare caused by light reflected off metal parts can also cause false detections. Lighting equipment can help reduce the effects of glare, but configuring the lighting conditions requires skill, and changeovers take a lot of time.
Vision sensors, however, can automatically configure settings for color differentiation, glare, and other factors. This means stable detection is possible at all times even without separate lighting equipment, and even subtle color differences are detected. The sensors are also able to automatically detect when the target changes to reconfigure detection conditions and lighting settings as necessary.

In a conventional system, performing various detections such as presence, contour, color, and area means installing color sensors, photoelectric sensors, or laser displacement sensors according to the desired detection.
Vision sensors, however, can be used to use different detection conditions at different locations all at once. The ability to perform detection using the various configured settings all at once means differentiation that conventionally required several sensors can be done with just one device. Also, the ability to save the detection conditions means they can be easily retrieved as needed. This helps significantly reduce not only the number of sensors needed but also the time required for changeovers.

Misaligned products can cause false detections with sensors that detect points on a target, such as photoelectric sensors. This means positioning devices are required to ensure accurate positioning.
Vision sensors detect the shape of a target by capturing the surface of the target as an image. The target’s features are then analyzed to determine the position and orientation of the detection target. This ensures accurate differentiation even if the position of the passing target is misaligned.
In addition, with the IV2 Series, the detection area is wide (550 × 412 mm 21.65" × 16.22"), and accurate detection and differentiation are possible anywhere within the area even for misaligned targets.

Photoelectric sensors and contact-type displacement sensors capture information from a specific point on a target, requiring the same number of sensors as detection points. All of these sensors must be installed and configured before performing detection. Moreover, if the product type changes, a significant amount of time is required for changeovers.
The IV2 Series, however, captures the surface of the target, enabling detection of up to 16 different points within the field of view. This includes not only presence checking and shape analysis, but also color, pitch, and various other differences. Up to 32 different detection setting configurations can also be saved (for 32 different products), and changing the settings is as easy as loading the configuration for the specific target.
This means the number of sensors required for difference checking as well as the time spent on installation, setup, and changeovers can be significantly reduced.

Glare caused by the strong reflected light from polished metal surfaces can cause false detection with conventional systems. With photoelectric sensors and color sensors, color is detected at a single point, making stable detection difficult.
The IV2 Series vision sensor with built-in AI, however, features automatic focus and brightness adjustment functions that enable stable detection by determining the optimal focus and brightness without separate lighting equipment even if glare is present. Detection settings for color differentiation, contours, and other factors are also configured automatically by the AI.
The IV2 Series uses AI to ensure differentiation even with glossy surfaces by determining the ideal detection conditions, allowing users to construct inspection systems with superior detection stability.