From: "Peter John Smith" 
To: 
Reply-To: "Peter John Smith" 


The distinction between resolving and detection has been pointed out by
others.  The difference may be quite staggering to those who have
considered nothing but the often quited 'resolution limits' in simplistic
cases only.

A few factors not mentioned.

It helps to know beforehand what is being detected.  In most cases it is
essential.  For example, if one knows there is one high contrast line,
detection becomes easier and an observer can reliably give the line
orientation. . Of course, if well under resolution limits, we cannot
confirm if is one or two lines.  And if it was one bright and one faint
line we are unaware of this.

If the object geometry is known, positional measurements may be made to a
resolution far far below the usually quoted resolution limits.

It is possible to quite accurately find the centre of a huge Airy disk and
ring pattern.  But, again, we are assuming a point source.  Since we do not
have the resolution to confirm the shape of the source, this preknowledge
is essential.

To achieve this apparent magic, we do not need super accurate optical
surfaces. Like to detect something 1/10 smaller than resolution size, we
dont need surfaces 10 times better.  Certainly, distinctly better than 1/4
wave wavefront is desireable but 1/40 is not needed.  Minimising surface
smoothness and scattered light is fairly important.



Peter Smith.

--- BBBS/NT v4.00 MP