home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!utcsri!torn!watserv2.uwaterloo.ca!watmath!watcgl!watpix.uwaterloo.ca!awpaeth
- From: awpaeth@watpix.uwaterloo.ca (Alan Wm Paeth)
- Subject: Re: Mercator Projection
- Message-ID: <BxH0uF.CBn@watcgl.uwaterloo.ca>
- Sender: news@watcgl.uwaterloo.ca (USENET News System)
- Organization: University of Waterloo
- References: <a34uTB4w165w@netlink.cts.com> <israel.721212129@unixg.ubc.ca> <1992Nov8.214329.27209@CSD-NewsHost.Stanford.EDU>
- Date: Mon, 9 Nov 1992 22:51:51 GMT
- Lines: 49
-
- In article <1992Nov8.214329.27209@CSD-NewsHost.Stanford.EDU> pratt@Sunburn.Stanford.EDU (Vaughan R. Pratt) writes:
-
- > [...questions on the derivation and use maps / conformal (complex) maps...]
-
- The Mercator projection has no simple geometric analog. Being cylindrical
- (meridians/lines of longitude map onto parallel, vertical lines on the chart)
- it is often incorrectly assumed to be developed from a sphere and tangent
- cylinder model.
-
- The mapping function is the integral of the secant, called (in my CRC) the
- inverse Gudermannian. Its most common formula is
-
- f(x) = ln (tan (x/2 + Pi/4))
-
- The Taylor expansion (about zero) of the function and its inverse have
- identical terms, though one contains alternating signs, the other all
- positive. This property can be used to explain the Gudermannian's use of
- interrelating trigonometric and hyperbolic formulae without the use of
- an imaginary unit. (I view the function as the real portion yanked from a
- complex analytical function and hence a conformal map). The series are both
- odd and resemble arcsin(x) and sin(x) through the cubic term.
-
- Conformal maps preserve angle measure (and their handedness). For instance, the
- Stereographic map is the image of the complex function f(z) = z^(-1). It is
- conformal and by convention has a proper orientation ("up" on the chart is
- "north" along the sphere) at its point of projection. But elsewhere, the
- ``compass rose'' must be rotated. Because the Mercator projection is
- cylindrial, "north" is upwards up along any meridian anywhere on the chart. So
- being both conformal and cylindrical are necessary and sufficient to product a
- chart on which compass heading may be plotted directly. This was Mercator's
- intent in his derivation (not his real name, BTW) and explains the chart's
- widespread nautical application.
-
- Because of the excessive scale distortions near the pole, transverse or
- oblique projections may rotate the pole onto the equator or another arbitrary
- position. Transverse Mercator projections have geodetic applications. Oblique
- projections may be used to produce long-haul give over-the-pole charts with
- little distortion along a select great circle route, though for general
- aeronautical use Lambert's (second) conformal conic is more widely used.
-
- Most of what I've written above is excerpted from my entry appearing in
- _Graphics Gems_ (Academic Press, 1990, Andrew Glassner, ed.), written with the
- scientist/mathematician in mind. The volume includes the derivation and
- accompanying figures for a number of common charts, including representative
- conformal and equal-area varieties.
-
- /Alan Paeth
- Computer Graphics Laboratory
- University of Waterloo
-