Research Document Imaging Image Processing Noise

Search

Noise Degradation in Document Imaging
Dr. Eugene Ageenko

Image digitization process.

The process to convert the documents into digital form includes always noise degradation that affects image quality and makes further image processing and analyzing more difficult.

1. Light reflected from the object pass throw the optical system, and form the image radiant energy distribution L.
2. Light energy intensity is transformed to the electrical current V by the sensor.
3. Electrical current is digitized by Analog-to-Digital Converter.

Noise degradation is originated from:

• low quality originals (e.g. old blue-prints),
• quantization errors in digitization process,
• non-optimal light and contrast settings of the copying/scanning process,
• document transmission errors,
• paper defects and/or dirty optics.

Noise degradation process.

Noise appears in the images in two forms: 

• additive - randomly scattered noise pixels;
• content-dependent - distorting the contours of printed objects (lines, characters) by making them ragged.

Additive noise can be efficiently removed using morphological filters, whereas the content-dependent noise is removable only by image smoothing, which degrades image resolution, or innovative Feature-based or Context-based filtering.

Examples of noise degradation.

Noise-free image:

Additive noise:

Content-dependent noise:

Composite noise:

Noise free and noisy image samples: