merge_correlation¶
- sofia_redux.instruments.forcast.merge_correlation.merge_correlation(data, header, variance=None, maxshift=999999999.0, normmap=None, upsample=100, maxregister=16, resize=True)[source]¶
Merge an image using a correlation algorithm
Add each frame of the data to a 2-d summation frame in a manner appropriate to the current reduction scheme, then average by the number of frames
- Parameters:
- datanumpy.ndarray
Data to be merged i.e. frame with target images (nrow, ncol)
- headerastropy.io.fits.header.Header, optional
FITS header of the new input data file
- maxshiftfloat, optional
Maximum possible value of the shift
- variancenumpy.ndarray, optional
Propagate provided variance. Must match shape of data array (nrow, ncol).
- normmapnumpy.ndarray, optional
Array to hold the normalization map
- upsampleint, optional
Determines the fractional pixel accuracy of the registration algorithm. An upsample factor of 100 will result in registration accurate to one 100th of a pixel.
- maxregisterint or float or array-like, optional
The maximum pixel shift allowed in each dimension applied during the registration algorithm. Order of dimensions is (x, y). The initial chop and nod estimates will be retrieved from the header. A maximum correlation will be searched for in the range (chop_or_nod +/- maxregister). Set to None to perform a maximum correlation search over the entire image.
- Returns:
- numpy.ndarray, numpy.ndarray
Merged image i.e. frame with images of object at chop and nod positions merged.