<img alt="" src="https://secure.leadforensics.com/11743.png" style="display:none;">

Choosing The Right Sensor For Your Machine Vision Application

When choosing a machine vision camera, there are certain parameters that can make a big impact on your application and are less known than the standard parameters such as resolution and framerate.

Specifications such as sensitivity, noise level and dynamic range are gaining more and more importance to solve critical applications. To allow users to compare these parameters between different manufacturers, the European standard EMVA1288 was created, ensuring that manufacturers report the results of their products following the standard and that the information is clearly accessible to the user.

With Sony’s discontinuation of CCD sensors and the strong positioning of the Sony Pregius sensor family, the EMVA1288 standard has helped to ensure a smooth transition from CCD to CMOS and help Machine Vision system manufacturers, who face difficult choices in the pursuit or the perfect industrial camera.

The image quality parameters measured by the standard and which are key when selecting a camera are as follows.

Quantum Efficiency (also known as QE):

Measures the percentage (%) of photons converted to electrons in the sensor for a given wavelength. A high percentage indicates a more efficient sensor converting the received light into an electrical signal, therefore easier to detect light. This feature is especially beneficial for applications with low light or ones that require high sensitivity at certain wavelengths, such as traffic monitoring applications, or scientific fluorescence applications. The percentage of quantum efficiency depends solely on the sensor.

Noise (also known as Temporal Dark Noise or Read Noise):

It is measured in Electrons (e-) and is the amount of noise in the sensor when there is no presence of signal. A low noise level translates into a cleaner image. All sensors introduce some noise, as this is caused by the electronics of the sensor and by the design of the camera. The new sensors use advanced techniques for noise reduction. Similarly, camera manufacturers can add modifications to the camera's firmware to further lower noise levels. Typical machine vision applications that can benefit from low noise cameras are ANPR, fruit sorting and life sciences.

Saturation capacity:

It is measured in Electrons (e-) and measures the amount of charge a pixel can handle. Each pixel behaves like a bucket that can contain electrons. The saturation capacity indicates the maximum number of electrons that each individual pixel can store and is generally related to the pixel size of the sensor. The saturation capacity affects the Dynamic Range; for example, a high saturation capacity increases the Dynamic Range potential. The saturation capacity depends only on the sensor in the camera.

Signal to Noise ratio:

It is measured in Decibels (dB) or in Bits. It is the ratio between the signal at saturation levels versus the noise at saturation levels. The higher the number, the better the image contrast and clarity will be retained. For machine vision applications with low light levels, a high signal-to-noise ratio will be key in achieving good results.

Dynamic Range:

It is measured in Decibels (dB) or in Bits, and provides the ratio between the signal at saturation levels and the minimum signal that the sensor can detect. The higher the number, the higher levels of detail can be captured. In other words, Dynamic Range describes the ability of the camera to detect the maximum and minimum levels of highlights and shadows, meaning that industrial cameras with the highest dynamic range are those that capture more detail, when there are large amounts of variability in illumination (brightness and darkness). Dynamic Range results depend on both the sensor and the design of the camera, as the camera's Analog-Digital converter can also affect it. Dynamic range is important for machine vision applications such as traffic, sports analysis, and broadcasting.

Sensitivity Threshold:

It is measured in Photons (γ) and refers to the number of photons required for the signal to equal the level of noise. The lower the number, the better the camera's ability to capture useful image data above noise. The Sensitivity Threshold provides a good measure to understand how the camera will perform in low light applications, as it takes into account other parameters such as Quantum Efficiency and Noise Levels of the camera (both dark noise and shot noise).

EMVA1288 data for our range of cameras is available on our website and it’s there to help you choose the best machine vision camera for your application.

If you need further assistance choosing the right sensor for your application, don't hesitate to contact us:

info@clearview-imaging.com
01844 217 270

Comments

MACHINE VISION GIVES COMPUTERS THE POWER TO SEE

Machine Vision lets computers read barcodes, data matrix codes, direct part marks, optical character recognition and optical character verification – Learn more about the computer vision technology here

Learn More
Laptop Image Small-1