| | 1 | (Up to [wiki:PsPhot psPhot] Notes) |
| | 2 | |
| | 3 | I have been running tests of psPhot on ISP images. Since this camera has such substantial PSF variations across the field, I have found it necessary to allow the PSF parameters to have more than linear variation terms. I have added code to allow the user to specify the desired PSF parameter polynomial coeffs by defining a polynomial in the psphot.config using the metadata config polynomial representation, associated with the keyword PSF.TREND.MASK. The defined polynomial identifies the polynomial order as well as the masked and unmasked coeffs. For ISP, at least, bumping the PSF variation to 2nd order helps with the ap-fit scatter (from perhaps 0.05 to about 0.02). |
| | 4 | |
| | 5 | I also found an error in the PGAUSS model flux integration (I cut-and-paste the functional form from the QGAUSS model; this resulted in massive ap-fit values until I tracked it down). |
| | 6 | |
| | 7 | An important remaining issue with psPhot is that it seems to be rejecting fitting objects too agressively. Too many objects are not subtracted from the image because their fits are failing (eg, 200/1600 in an image of the triffid nebula). This are not obviously terrible objects, and it may be due to the rules on the chisq limits allowed for a successful fit. If so, I need a better way to set the limits for reasonable data. |
| | 8 | |
| | 9 | --EAM @ 2006.10.13,14:40 |
