Plot_all_matches

We have work on the code of plot_all_matches.m. This code allows us to obtain the detection rate, rms error and SNR values. I have a table. I am posting the figures that determine in this case (13 mHz using the MAAS trader container ship) which is the best method.

The first figure shows the same FFT for both methods and we are only changing the threshold.

The second figure is a comparison between both methods.

Error Bars

The document “Ship Detection at Codar” states on page 38 “Each instance is
represented by a square (error box) with the half height as the uncertainty (σ by (14)) and
half length as the half FFT window”

So the full length is represented by the full length of the FFT?  Reference 14 is a private communication with Jimmy Isaccson.  Jimmy, can comment on how the standard deviation is calculated?