CURRENT ISSUE
Contests
Test your eq
|
|
Issue #218 September 2008
Problem 3—When processing video signals in the digital domain, sometimes you need a line buffer and sometimes you need a frame buffer. What’s the difference, and why would you need one versus the other?
Answer 3—Some video processing operations work on one pixel at a time, while others work on a “neighborhood” or “kernel” of pixels at a time. The latter group includes spatial filters and feature-extraction algorithms such as edge detectors. A line buffer, which contains data from several adjacent scan lines, is used to make data about a particular group of pixels (e.g., a 5 × 5 kernel) available together for processing.
Other operations work on pixels in the time domain—the successive values of a single pixel from one frame to the next are important. For these operations, a buffer that can hold the pixel data for one or more complete frames is required. Examples include time-domain filtering for noise and handling the transitions among different frame rates within a system.
Frame buffers can also be used to hold additional pixel-by-plxel data that is combined in some way with the image data. Examples include compensating for image sensor irregularities by storing individual scale and offset values, or holding computer-generated text and imagery that is to be overlaid on the video signal.
Back to Issue #218 Questions | Test Your EQ Archive List
|
