| filter.discardNegative | bool | true, false | true | If set to `false` then SoFiA will not remove detections with negative flux. Note that reliability filtering must be disabled for negative sources to be retained. Also note that negative sources will not appear in moment 1 and 2 maps. This setting should only ever be disabled for testing or debugging purposes, but never in production mode. |
| filter.discardNegative | bool | true, false | true | If set to `true` then SoFiA will not remove detections with negative flux. Note that reliability filtering must be disabled for negative sources to be retained. Also note that negative sources will not appear in moment 1 and 2 maps. This setting should only ever be disabled for testing or debugging purposes, but never in production mode. |
| filter.minPixels | int | ≥ 0 | 0 | Minimum total number of spatial and spectral pixels within the source mask for detections to be considered reliable. Detections with fewer pixels will be removed from the catalogue. |
| filter.minSNR | float | ≥ 0.0 | 3.0 | Lower signal-to-noise limit for reliable sources. Detections that fall below this threshold will be deemed unreliable and removed from the catalogue. The value denotes the integrated signal-to-noise ratio, SNR = F_sum / \[RMS * sqrt(N * Omega)\], of the source, where Omega is the solid angle (in pixels) of the point spread function of the data, N is the number of spatial and spectral pixels of the source, F_sum is the summed flux density and RMS is the local RMS noise level (assumed to be constant). Note that the spectral resolution is assumed to be equal to the channel width. If `BMAJ` and `BMIN` are not defined in the input FITS file header then Omega will be set to 1 by default, thus assuming a beam size of 1 pixel. |
| filter.minSNR | float | ≥ 0.0 | 3.0 | Lower signal-to-noise limit for reliable sources. Detections that fall below this threshold will be deemed unreliable and removed from the catalogue. The value denotes the integrated signal-to-noise ratio, SNR = \|F_sum\| / \[RMS * sqrt(N * Omega)\], of the source, where Omega is the solid angle (in pixels) of the point spread function of the data, N is the number of spatial and spectral pixels of the source, F_sum is the summed flux density and RMS is the local RMS noise level (assumed to be constant). Note that the spectral resolution is assumed to be equal to the channel width. If `BMAJ` and `BMIN` are not defined in the input FITS file header then Omega will be set to 1 by default, thus assuming a beam size of 1 pixel. |
