This tool allows the user to quickly compute the WCS solution
for an LRS pre image. Below is an example of a solution
for an observation of an SDSS candidate for spectroscopic
followup. We use a DSS astrometric solution to draw an
(approximately) 6-sigma error circle for the source.
The LRSVIEW routine was used to pre-process the LRS image
and remove the strong radial vignetting function across the
field. A low-pass filter was constucted using a 12-pixel-wide
running mean filter. This image was subtracted from the
original to yield a map of the high spatial frequency
image structure. Such images are much easier to use with
finding charts.
Source position: 11:*6:07.41 +11:*1:25.7 The astromteric solution results: GUI-acquired rough source positions: Xm Ym Xwcs Ywcs RA_2000 DEC_2000 386.3 290.2 108.6 176.1 11:*6:10.18 +11:*1:52.3 258.8 160.2 192.6 166.0 11:*6:04.42 +11:*1:42.5 274.0 363.1 113.6 113.6 11:*6:09.82 +11:*0:49.4 277.5 460.9 75.8 89.0 11:*6:12.41 +11:*0:24.5 537.1 287.9 66.9 233.0 11:*6:13.06 +11:*2:49.4 505.9 348.3 53.7 205.2 11:*6:13.96 +11:*2:21.4 428.3 360.3 70.7 173.0 11:*6:12.78 +11:*1:49.0 RMS(Ra,Dec): 0.2346 0.5642 n = 7
Figure 1: Using WCS from the above astrometric solution we can locate the known position of our LRS target. The purple circle above is a 6-sigma error region. Only one well-detected source (30 sec LRS) lies inside this area. |
---|
The majority of LRS fields would contain enough objects easily visible in a DSS image to provide a WCS solution good to a few tenths of an arcsec. If users provide RA,DEC in some known astrometric system, then some variant of CALIM could be used to greatly aid in the acquisition of faint LRS targets. The current approach of using paper finding charts, produced in a wide variety ways and qualities, should really become a thing of the past.
Having derived a suitable astrometric solution, CALIM will next attempt to derive a photometric zeropoint for the image. Standard star magnitudes are read from a file of standard field photometry and matched up to the image sources via the new WCS. At present, the HET standard field set consists of BVR photometry of N7006, N7790, and N4147 (Odewahn etal 1993, PASP) and BVRI photometry of M92 (Christian etal 1985, PASP). For each source a set of aperture photometry data are computed and used to derive a photometric zeropoint. In addition, the local sky surface brightness is computed for each source. Some examples of figures generated with this portion of CALIM are shown below in Figures 2 through 4.
Figure 2: An LRS V image of NGC 4147. The position of each standard source is used as a starting-point in measuring photometry on the LRS image. After sky determination, intensity weighted moments are used to computed the position, size, shape and orientation of each source. Magnitudes are derived with a user-specified aperture size. The sizes of the source markers in the image above reflect the size of each image determined from the intensity-weighted second moment. |
---|
Figure 3: The V-band sky surface brightness (mag/sq.arcec) for each source measured in Figure 2. The sky brightness is slightly correlated with the source brightness. These sky measures are very bright due to the fact that a 90% illuminated moon was 75% away from NGC 4147 at the time of the observation. |
---|
Figure 4: The individual ZP estimated from the sources shown in Figure 2 as a function of distance from the LRS image center (in pixel units). Note the strong trend in ZP with radial distance caused by the strong vignetting function in an LRS image. Points, such as those encased in yellow circles above, may be isolated by the user for rejection prior to a linear regression. |
---|
A number of refinements have been made to CALIM to allow the use of general HET LRS fields. Some initial work with this upgraded code to investigate rho position errors is summarized below.
HET Rho Measurements The following plate solutions for LRS fields were derived with CALIM using DSS images for an intial solution, and USNOA2.0 astrometry for the final least squares fits. The important column below is labeled DEL and is the difference between the HET (LRS header) PA (position angle) on the sky and the value for the same quantity I derive in the fit. In general, these numbers are low. The one high case (***) occurred when we were making changes to the rho zeropoint. I believe these results show that we can use CALIM to monitor rho positioning periodically. This method is not as accurate as the iraf method that Matt runs. That method uses bright cluster stars with good astrometry. I use USNO astrometry based on Schmidt plates, hence we see rms residuals on the order of a few tenths. However, we can use this method anywhere we take an image with LRS! # scale DEL dRA dDEC skyPA_het date RA Object S3 0.4670 0.03 0.276 0.290 242.2 20040504 SO N4147 S6 0.4680 0.11 0.464 0.575 96.3 20040811 S0 J2025+3343 S5 0.4700 0.03 0.727 0.370 220.7 20040811 SO J2120+0533 --------------------- PFIP Take-Down -------------------------------- S7 0.4717 -0.15 0.471 0.576 134.6 20040904 MS N7790 S10 0.4671 0.29 0.313 0.441 175.8 20041010 SO Munich S9 0.4683 0.60 0.458 0.663 -45.4 20041016 MS N7790 S11 0.4705 0.97 0.360 0.455 175.0 20041018 BR Munich (***) --------------------- Final JRF Rho Fix ----------------------------- S8 0.4670 -0.08 0.328 0.559 318.6 20041019 BR N7790 S12 0.4686 0.23 0.477 0.370 73.0 20041020 BR J0742+4900 scale = derived pixel size in arcsec DEL = skyPA_het - skyPA_fit (in units of degrees) dRA,dDEC = rms residuals (arcsec) in RA,DEC skyPA_het = position angle of LRS column on sky (from LRS header info) (***) I did this fit many ways and continually came up with a 1 degree offset. I was concerned because I thought the mask was aligned and Jim's rho fix was in place. Neither was true! I ran MOSOBS on the setup image and found approximately a 1 degree rotation needed to align the mask! Also, this image was taken on 20041018 (which is a day before JF's final fix). Note that my 20041010 Munich mask image gives 0.3 degree agreement between HET and fitted skyPA. This was about a week prior to the rho fix, and I am surprised the agreement was so go. I had to rotate the field by 0.8 degrees (from MOSOBS) to align the mask. Hence, what HET thought was at PA=175 had to be placed at 175.8. What is confusing to me is that my plate solution gives a close match to that at PA=175.5. Note that MS's 20041016 N7790 was also off by 0.6 degree, and this was what motivated the rho fix by JF. The initial fix was made in the wrong direction, and this accounts for the large shift needed for BR's 20041018 image. Once correct the rho sign was used (after 20041018), we see that the later fields have HET and fitted sky PAs that agree well.