From a Ph.D. thesis by Roland Schwarz.
The first generation of scanners for bathymetry came into use shortly after the invention of the LASER. Although in the beginning mainly only analog electronics was available for the evaluation of the signals, later the entire trace of the backscattered echo was recorded digitally. The advent of affordable computers finally opened the way to more complex signal processing.
For the determination of the elevation of the underwater bottom it is necessary to identify two significant time instants in the waveform. The first is when the light impulse enters the water and the second is when it hits the bottom. It is especially important to know the first moment, because from this moment on the impulse moves slower and in a different direction.
The standard method to identify an instant of time in a signal is by Gaussian decomposition of the signal. Underwater, however, the method suffers from the problem that a lot of distributed small particles cause clutter that is hindering the exact decomposition. For a tenuous distribution of such particles, the waveform is of exponential character.
In this thesis I therefore introduce a model consisting of exponential segments that describe the effect of the particles and Dirac shaped pulses that describe the effect of discretely located scatterers. This description is however not sufficient yet to account for the received signal form. The exponential model has to be convolved with the system waveform to yield a correct representation of the received signal. By minimizing the difference of this representation and the measured data, the parameters of the exponential model can be retrieved.
I present a procedure, which I call exponential decomposition, by which the actual processing can be done. The effectiveness of the procedure is verified on the basis of data collected in a tributary of the Danube river. The correctness of the results is confirmed using GNSS surveyed control points.
An important aspect for the modeling of signals is that the model is physically correct. An underestimated effect in laser bathymetry is that pulsed light propagates more slowly than conventionally assumed. Since the effect in the context of laser bathymetry has not yet been discussed, I describe an experiment I performed that confirms the effect in its predicted magnitude.
Furthermore, I deal with the questions whether a single wavelength system is feasible and what the smallest measurable depth in laser bathymetry is. I describe an experiment I performed that confirms the effect in its predicted magnitude.
Furthermore, I deal with the questions whether a single wavelength system is feasible and what the smallest measurable depth in laser bathymetry is. I describe an experiment I performed that confirms the effect in its predicted magnitude.
For the complete paper on analysis and processing of bathymetric lidar signals CLICK HERE.
