as know, there functions in cmos image sensor isp (image signal processor). specifically, i'd know difference between binning , sub-sampling. think these purpose same reduce image size. however, i'm not sure why these functions exist? purpose?
binning , sub-sampling reduce image size have suspected, focus on different things. let's tackle each issue separately
binning
binning in image processing deals quantization. closest thing can think of related known data binning. basically, consider breaking image distinct (non-overlapping) m x n tiles, m , n rows , columns of tile , m , n should smaller rows , columns of image.
if consider grid of m x n pixels, of these pixels replaced representative colour. way representative colour calculated done in many ways... average popular method. reason why binning performed data pre-processing technique used reduce effects of minor observation errors. reduces amount of information representative of image, , reduces image size reducing amount of unique colours represent image.
in addition, binning data may reduce impact of noise impacts cmos sensor on final processed image, @ cost of lower dynamic range of colours.
sub-sampling
sub-sampling in case of image processing deals image resizing. it's called image scaling. goal take image , reduce dimensions smaller image result. binning deals keeping image same size (i.e. same dimensions original) while reducing amount of colours reduces amount of space image takes up. subsampling reduces image size removing information together. when subsample, interpolate or smooth image reduce aliasing.
sub-sampling has application in video processing - in mpeg video encoded in ycbcr. y luminance while cb , cr chrominance pairs. tend notice changes in luminance rather chrominance, , chrominance subsampled reduce amount of space taken video. specifically, the human visual system has poor acuity when comes colour information luminance / intensity. usually, chrominance values filtered subsampled 1/2 or 1/4 of of intensity. rather high subsampling rate, don't notice differences in terms of perceived image quality.
this rather rough introduction on differences between them both, hope gives enough of you're after purposes.
good luck!
Comments
Post a Comment