Defect Pixel Correction Even More Critical for High-Resolution Machine Vision Cameras

Posted by Gretchen Alper on Wed, Jul 18, 2012

When there are pixels on the image sensor whose response is an extreme outlier, these defect pixels can be corrected to ensure that they do not affect the precision or accuracy of the system.

With the introduction of large CMOS image sensors and cameras on the market the topic of defect pixel correction is becoming more important. With more than 25 million pixels or even 1 million pixels on an image sensor, not every single pixel can perform perfectly.  Inspection and measurement systems require uniform, undistorted, consistent images despite imperfect components.  There are several techniques for camera manufacturers to achieve this, and sometimes there are pixels whose response differs so significantly from the mean response of all of the other pixels that they need to be excluded entirely.  These pixels are referred to as defect or defective pixels.

There are 2 types of defect pixels:

  1. Hot pixels:  Pixels with an extremely large offset for example due to excessive dark current (local impurities/lattice defects).
  2. Dead pixels:  pixels with much lower sensitivity
    • Show as white spots in images.
    • Depend on sensor parameters such as temperature and integration time.
    • Show as darker spots with respect to the surrounding pixels.

Adjacent pixel defects are called cluster defects.

Higher temperature operations and long integration times can result in a greater number of defect pixels.   

The camera manufacturer can take the image sensor and implement defect pixel correction combined with other functionality to create the uniform images required.  

For certain applications, such as semiconductor wafer metrology or optical CD metrology, defects in certain positions on the sensor can impact the quality of the overall system.  When sub-pixel accuracy is required, each pixel response needs to be known before any measurements are made.  Without proper control of defect pixels, the accuracy of the measurement is affected.

Special attention to defect pixels is increasingly important with the move to higher resolution area scan image sensors and subsequent cameras as the number of defect pixels also increases.  If the specification for a 4 Megapixel image sensor were a maximum of 200 defect pixels, this same specification would translate to up to 800 defect pixels for a 16 Megapixel image sensor.


For more information on how camera manufacturers create a uniform image, read our free ePaper below.

compare machine vision cameras

Topics: Image Quality Improvements

Previous blog:

Next blog: