Detection of weak and overlapping pulses from waveform airborne laser scanning data. SilviLaser 2008

Airborne laser scanning has become a popular technique to estimate canopy height and forest structure. However, current discrete laser scanning systems still suffer the limitation of 3m multi-target resolution and the loss of information about range estimation. In order to improve range resolution and accuracy, new generation small-footprint waveform laser scanning data were investigated. A new approach was developed to detect targets from complex overlapping and weak pulses, which are likely to occur in vegetated areas. The algorithm is based on the popular Gaussian decomposition method and contains two main processing procedures. The shapes of overlapping pulses are analysed to find visible peaks and overlapping peaks, and then reasonable constraints and checks are applied to the fitting process. The test results showed the developed detection algorithm resolved overlapping pulses very well as long as the pulse shapes illustrated asymmetric behaviour or a non-Gaussian distribution. Weak pulses exhibiting a Gaussian shape were also successfully detected