Some Notes On Data Smoothing

The topic of data filtering can be pretty obscure for the general public.  Taking raw data and smoothing out the lumps can seem like deliberate falsification.  It isn't.

In the early days of space exploration it was only possible to produce images of fairly low resolution.  Special techniques were used to enhance the images.

In the first image, below, I have colorised a block of 9 pixels to demonstrate a simple set of image enhancement techniques.  Each pixel in an image is compared to its immediate neighbors and an operation is performed to filter the picture.  It could hardly be simpler.

Here are some pictures to illustrate a few of the many effects that can be obtained by simple filtering of the raw data.

If you squint at the top picture it looks more like the bottom one.  The human eye-brain system is a natural raw data filter.

Here are some 'special effects' easily achieved with basic maths.  The sign below, which translates more usually as 'mind your head', is a portion of a larger image.

Even the last image retains the core information which is of interest to the human observer.

Much has been made about various interpretations of various climate data sets.  My purpose in showing these images is to demonstrate that even the last image shows the data of greatest interest to a linguist.  If a linguist uses such data to recover previously unreadable text as part of a series showing how language has changed down the centuries, there are no 'language change' deniers ready to pounce on the FACT that all the green bits have been eliminated by a dishonest scientist to bias the result.  Why is it that what applies to the maths of linguistics is not seen to apply just as honestly to climate science?

More information about image manipulation, and code:

http://www.codeproject.com/KB/GDI-plus/csharpfilters.aspx

Once you have read the information at that link you will know how this old fossil made his old fossil avatar. :-)