Use fit instead of RMS for prealignment
Currently, the Prealignment module uses the RMS to adjust the telescope planes such that the correlation peak is located around 0. In case of a skewed background this can cause problems since the location of the peak cannot be determined easily from the RMS.
Suggestion: Find the max. value of the histogram and perform a (Gaussian?) fit around it to obtain the peak position. Alternatively, a proper peak finding algorithm can be implemented.