Feature analysis of LIDAR waveforms from forest canopies

Airborne light detection and ranging (LIDAR) can detect the three-dimensional structure of forest canopies by transmitting laser pulses and receiving returned waveforms which contain backscatter from branches and leaves at different heights.We established a solid scatterer model to explain the widened durations found in analyzing the relationship between laser pulses and forest canopies,and obtained the corresponding rule between laser pulse duration and scatterer depth.Based on returned waveform characteristics,scatterers were classified into three types:simple,solid and complex.We developed single-peak derivative and multiple-peak derivative analysis methods to retrieve waveform features and discriminate between scatterer types.Solid scatterer simulations showed that the returned waveforms were widened as scatterer depth increased,and as space between sub-scatterers increased the returned waveforms developed two peaks which subsequently developed into two separate sub-waveforms.There were slight differences between the durations of simulated and measured waveforms.LIDAR waveform data are able to describe the backscatter characteristics of forest canopies,and have potential to improve the estimation accuracy of forest parameters.