Considerations with ultra high resolution (12 MP+) machine vision cameras

For many applications, machine vision cameras with increased resolution can offer advantages such as increased field of view or greater accuracy.  Now ultra high-resolution cameras (greater than 12 Megapixels) are available with acceptable frame rates such as 12 Megapixels at 66 fps.  With this change, there are some considerations…

Changing sensor technology

When increasing resolution beyond 12 Megapixels, it can also involve a switch in sensor technology as very high-resolution CCD cameras have limited frame rates.  The latest generation CMOS sensors eliminate the trade off for very high resolution and fast frame rates, with for example 12 Megapixels at 66 fps and higher.

With the changes from CCD to CMOS, there are particular image artifacts and other differences.  For some more detailed information: 

CCD vs CMOS Image Artifacts to Consider with CMOS Image Sensors

CCD vs CMOS Entering the Slope of Enlightenment

New Interfaces

The other change with higher resolutions at high speeds (with the trend towards

Even higher speeds) results in the need to transmit a lot of data.  Several new interface standards are now available to go beyond previously established standards.  Some are still not fully capable of these very large data rates, for example USB3 Vision is limited to 12 Megapixels up to 30 fps.

For more details about the pros and cons of different interfaces:

Considerations with CoaXPress, Camera Link HS, 10 GigE Vision, and USB3 Vision

Defect Pixels

More defect pixels with ultra high resolutions can be expected than what was familiar with 1, 2, or even 4 Megapixel sensors/cameras.  There are just more pixels so the number of defect pixels will go up.  Also, as the resolution of the sensor increases, the size of each sensor increases, limiting the number of sensors per wafer in production.  To then further limit the yield with very stringent blemish specifications would result in prohibitive cost increases.  More defect pixels and even cluster defects and row and column defects can be unavoidable.  

Cluster defects and/or row and column defects provide a challenge with defect pixel correction, as an effective correction requires the right context in the area of the defect pixel.  Some ultra high-resolution cameras include 2 directional defect pixel corrections

System Considerations

With larger resolutions, and subsequent sensor size results in the need for more complex optics and a different light source.

The trend towards keeping cameras very compact is in contrast with keeping the power consumption low.  Very small cameras may require a system design such that there is air flow to keep the camera cool.  Some camera manufacturers have found a balance, such as the Adimec Sapphire S-25A30 which offers 25 Megapixels at 32 fps with only about 6 W of power consumption.

Whether you require higher field of view, or other options that can be enabled through a region of interest function, then great ultra high-resolution cameras are worth considering.