7.5 Angular Resolution and Non-Random Scan Directions

An important issue is also how a non-random set of scan directions, in which some position angles appear much more frequently than others, could affect the general appearance, and particularly the angular resolution, of flux maps. In particular, since a sample of $6\times4$ pixels is rectangular with major side in the across-scan direction, the presence of a preferred scan direction in principle implies the elongation of the effective PSF, and therefore the loss of resolution, perpendicularly to this direction. Obviously the problem increases in size with the PSF asymmetry, and Figure 7.5 shows that for a sample size of $6\times4$ pixels its effects are essentially negligible in the case of scan directions randomly distributed in the intervals $[0^\circ,45^\circ] \cup [90^\circ,135^\circ]$ or $[0^\circ,90^\circ]$, but not so when the interval is reduced to $[0^\circ,45^\circ]$, in which case the two HII regions show an increased elongation. This latter case is however an extreme one, which is not likely to occur in practice.

Figure 7.5: Non-random scan directions and angular resolution of GAIA BBP flux maps. The two images show the same sky region as in Figures 7.3 and 7.4, but the flux maps are here reconstructed from 50 simulated observations with non-random scan directions. The sample size is $6\times4$ pixels in all cases but the scan directions are randomly distributed in the intervals: $[0^\circ,360^\circ]$ and $[0^\circ,45^\circ] \cup [90^\circ,135^\circ]$ for the upper row; $[0^\circ,90^\circ]$ and $[0^\circ,45^\circ]$ for the lower row. Only in the last case the image distortion with respect to the upper left image is not negligible. The side of each image is about 2 arcsec.