//#--------------------------------------------------------------------------- //# pks_maths.cc: Mathematical functions for Parkes single-dish data reduction //#--------------------------------------------------------------------------- //# livedata - processing pipeline for single-dish, multibeam spectral data. //# Copyright (C) 2004-2009, Australia Telescope National Facility, CSIRO //# //# This file is part of livedata. //# //# livedata is free software: you can redistribute it and/or modify it under //# the terms of the GNU General Public License as published by the Free //# Software Foundation, either version 3 of the License, or (at your option) //# any later version. //# //# livedata is distributed in the hope that it will be useful, but WITHOUT //# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or //# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for //# more details. //# //# You should have received a copy of the GNU General Public License along //# with livedata. If not, see . //# //# Correspondence concerning livedata may be directed to: //# Internet email: mcalabre@atnf.csiro.au //# Postal address: Dr. Mark Calabretta //# Australia Telescope National Facility, CSIRO //# PO Box 76 //# Epping NSW 1710 //# AUSTRALIA //# //# http://www.atnf.csiro.au/computing/software/livedata.html //# $Id: pks_maths.cc,v 1.7 2009-09-29 07:45:02 cal103 Exp $ //#--------------------------------------------------------------------------- //# Original: 2004/07/16 Mark Calabretta //#--------------------------------------------------------------------------- // AIPS++ includes. #include #include #include #include #include #include // Parkes includes. #include //----------------------------------------------------------------------- nint // Nearest integral value; halfway cases are rounded to the integral value // larger in value. No check is made for integer overflow. Int nint(Double v) { return Int(floor(v + 0.5)); } //---------------------------------------------------------------------- anint // Nearest integral value; halfway cases are rounded to the integral value // larger in value. Double anint(Double v) { return floor(v + 0.5); } //---------------------------------------------------------------------- round // Round value v to the nearest integral multiple of precision p. Double round(Double v, Double p) { return p * floor(v/p + 0.5); } //--------------------------------------------------------------------- median // Compute the weighted median value of an array. Float median(const Vector &v, const Vector &wgt) { uInt nElem = v.nelements(); if (nElem == 0) return 0.0f; // Generate the sort index. Vector sortindex(nElem); GenSortIndirect::sort(sortindex, v); // Find the middle weight. Float wgt_2 = sum(wgt)/2.0f; // Find the corresponding vector element. Float weight = 0.0f; Float accwgt = 0.0f; uInt j1 = 0; uInt j2; for (j2 = 0; j2 < nElem; j2++) { weight = wgt(sortindex(j2)); if (weight == 0.0f) { // Ignore zero-weight data; continue; } // The accumulated weight. accwgt += weight; if (accwgt <= wgt_2) { // Keep looping. j1 = j2; } else { break; } } // Compute weighted median. Float v1 = v(sortindex(j1)); Float v2 = v(sortindex(j2)); // Compute pro-rata value from below. Float dw = wgt_2 - (accwgt - weight); v1 += (v2 - v1) * dw / weight; // Find next non-zero-weight value. for (j2++ ; j2 < nElem; j2++) { weight = wgt(sortindex(j2)); if (weight != 0.0f) { break; } } if (j2 < nElem) { // Compute pro-rata value from above. Float v3 = v(sortindex(j2)); v2 += (v3 - v2) * dw / weight; } return (v1 + v2)/2.0f; } //---------------------------------------------------------------- angularDist // Determine the angular distance between two directions (angles in radians). Double angularDist(Double lng0, Double lat0, Double lng, Double lat) { Double costheta = sin(lat0)*sin(lat) + cos(lat0)*cos(lat)*cos(lng0-lng); return acos(costheta); } //--------------------------------------------------------------------- distPA void distPA(Double lng0, Double lat0, Double lng, Double lat, Double &dist, Double &pa) // Determine the generalized position angle of the field point (lng,lat) from // the reference point (lng0,lat0) and the angular distance between them // (angles in radians). { // Euler angles which rotate the coordinate frame so that (lng0,lat0) is // at the pole of the new system, with the pole of the old system at zero // longitude in the new. Double phi0 = C::pi_2 + lng0; Double theta = C::pi_2 - lat0; Double phi = -C::pi_2; // Rotate the field point to the new system. Double alpha, beta; eulerx(lng, lat, phi0, theta, phi, alpha, beta); dist = C::pi_2 - beta; pa = -alpha; if (pa < -C::pi) pa = pa + C::_2pi; } //--------------------------------------------------------------------- eulerx void eulerx(Double lng0, Double lat0, Double phi0, Double theta, Double phi, Double &lng1, Double &lat1) // Applies the Euler angle based transformation of spherical coordinates. // // phi0 Longitude of the ascending node in the old system, radians. The // ascending node is the point of intersection of the equators of // the two systems such that the equator of the new system crosses // from south to north as viewed in the old system. // // theta Angle between the poles of the two systems, radians. THETA is // positive for a positive rotation about the ascending node. // // phi Longitude of the ascending node in the new system, radians. { // Compute intermediaries. Double lng0p = lng0 - phi0; Double slng0p = sin(lng0p); Double clng0p = cos(lng0p); Double slat0 = sin(lat0); Double clat0 = cos(lat0); Double ctheta = cos(theta); Double stheta = sin(theta); Double x = clat0*clng0p; Double y = clat0*slng0p*ctheta + slat0*stheta; // Longitude in the new system. if (x != 0.0 || y != 0.0) { lng1 = phi + atan2(y, x); } else { // Longitude at the poles in the new system is consistent with that // specified in the old system. lng1 = phi + lng0p; } lng1 = fmod(lng1, C::_2pi); if (lng1 < 0.0) lng1 += C::_2pi; lat1 = asin(slat0*ctheta - clat0*stheta*slng0p); } //------------------------------------------------------------------------ sol // Low precision coordinates of the Sun (accurate to 1 arcmin between 1800 and // 2200) from http://aa.usno.navy.mil/faq/docs/SunApprox.html matches closely // that in the Astronomical Almanac. void sol(Double mjd, Double &elng, Double &ra, Double &dec) { Double d2r = C::pi/180.0; // Number of days since J2000.0. Double d = mjd - 51544.5; // Mean longitude and mean anomaly of the Sun (deg). Double L = 280.459 + 0.98564736*d; Double g = 357.529 + 0.98560028*d; // Apparent ecliptic longitude corrected for aberration (deg). g *= d2r; elng = L + 1.915*sin(g) + 0.020*sin(g+g); elng = fmod(elng, 360.0); if (elng < 0.0) elng += 360.0; // Obliquity of the ecliptic (deg). Double epsilon = 23.439 - 0.00000036*d; // Transform ecliptic to equatorial coordinates. elng *= d2r; epsilon *= d2r; ra = atan2(cos(epsilon)*sin(elng), cos(elng)); dec = asin(sin(epsilon)*sin(elng)); if (ra < 0.0) ra += C::_2pi; } //------------------------------------------------------------------------ gst // Greenwich mean sidereal time, and low precision Greenwich apparent sidereal // time, both in radian, from http://aa.usno.navy.mil/faq/docs/GAST.html. UT1 // is given in MJD form. void gst(Double ut1, Double &gmst, Double &gast) { Double d2r = C::pi/180.0; Double d = ut1 - 51544.5; Double d0 = int(ut1) - 51544.5; Double h = 24.0*(d - d0); Double t = d / 35625.0; // GMST (hr). gmst = 6.697374558 + 0.06570982441908*d0 + 1.00273790935*h + 0.000026*t*t; gmst = fmod(gmst, 24.0); // Longitude of the ascending node of the Moon (deg). Double Omega = 125.04 - 0.052954*d; // Mean Longitude of the Sun (deg). Double L = 280.47 + 0.98565*d; // Obliquity of the ecliptic (deg). Double epsilon = 23.4393 - 0.0000004*d; // Approximate nutation in longitude (hr). Double dpsi = -0.000319*sin(Omega*d2r) - 0.000024*sin((L+L)*d2r); // Equation of the equinoxes (hr). Double eqeq = dpsi*cos(epsilon*d2r); // GAST (hr). gast = gmst + eqeq; gast = fmod(gast, 24.0); // Convert to radian. gmst *= C::pi/12.0; gast *= C::pi/12.0; } //----------------------------------------------------------------------- azel // Convert (ra,dec) to (az,el). Position as a Cartesian triplet in m, UT1 in // MJD form, and all angles in radian. void azel(const Vector position, Double ut1, Double ra, Double dec, Double &az, Double &el) { // Get geocentric longitude and latitude (rad). Double x = position(0); Double y = position(1); Double z = position(2); Double r = sqrt(x*x + y*y + z*z); Double lng = atan2(y, x); Double lat = asin(z/r); // Get GAST (rad). Double gast, gmst; gst(ut1, gmst, gast); // Local hour angle (rad). Double ha = (gast + lng) - ra; // Azimuth and elevation (rad). az = atan2(cos(dec)*sin(ha), cos(dec)*sin(lat)*cos(ha) - sin(dec)*cos(lat)); if (az < 0.0) az += C::_2pi; el = asin(sin(dec)*sin(lat) + cos(dec)*cos(lat)*cos(ha)); } //---------------------------------------------------------------------- solel // Compute the Solar elevation using the above functions. Double solel(const Vector position, Double ut1) { Double az, dec, el, elng, gast, gmst, ra; sol(ut1, elng, ra, dec); gst(ut1, gmst, gast); azel(position, ut1, ra, dec, az, el); return el; }