Changeset 24973 for trunk/Ohana/src/addstar/doc/notes.txt

Timestamp:

Aug 2, 2009, 4:27:26 PM (17 years ago)

Author:

eugene

Message:

allow dvo to accept supplied zero point information

File:

• trunk/Ohana/src/addstar/doc/notes.txt (modified) (1 diff)

trunk/Ohana/src/addstar/doc/notes.txt

@@ +1 @@
+
+2009.08.02
+
+  ZERO POINTS: addstar currently has a couple of inconsistent ways that it 
+  measures or uses image zero points:
+
+  * the standard zero point: data in the database is saved as M_inst +
+    25.0.  This value is written to the image header and checked when
+    the image is loaded.  This is completely archaic, but removing
+    this concept from the DVO system will require some work and some
+    care.  This value is NOT reflected in the image->Mcal values
+
+  * on-the-fly calibration: addstar was used by CFHT skyprobe to
+    measure the sky transparency.  This calculation was done by
+    measuring the zero point between the Tycho stars in the database
+    and the stars in the image.  Star data is saved in
+    AddToCalibration as they are found (in find_matches, or
+    equivalent).  At the end of the match processing, the zero point
+    for the image was calculated with FindCalibration and saved to the
+    image table but NOT the measure table.  For skyprobe, the stars
+    are normally not saved to the database -- the Tycho stars are kept
+    in the measure & average tables, and used as the reference for the
+    zero point calculation above.
+
+  For Pan-STARRS, the zero point analysis is performed by psastro
+  before we run addstar.  However, there are a few ways we could
+  choose to apply the zero point calculation:
+
+  * use the per-chip zero point reported in the individual chip
+    headers (easy, but fairly wrong)
+
+  * calculate the per-exposure zero point from the chip headers and
+    supply.
+
+  * calculate the per-exposure zero point from the MATCHED_REFS table
+
+  We also need to consider the zero point error.  We currently have
+  only dM (poisson error) + dMcal (calibration error).  We need to
+  divide dMsys into dMcal (photometry calibration error) and dMap
+  (aperture correction error).
+
+  * zero point options:
+    * NOMINAL : set Mcal = 0.0
+    * CHIP_HEADER : each chip has zero point in header
+    * PHU_HEADER : a single exposure-wide value in the header
+    * CHIP_AVERAGE : determine zero point for exposure from chip zero points
+    * MATCHED_REFS : recalculate from table
 
 2009.02.11