Improved waveform reconstruction and parameter accuracy retrieval for hyperspectral lidar data

In this paper, we aim to provide a solution for choosing an optimized digitization frequency for full-waveform reconstruction of narrow FWHM laser pulses to improve the performance accuracy required for recognition and characterization of a wide range of materials with distinctive reflectance features. This type of analysis, for the first time, to the best of our knowledge, gives an assessment of the absolute accuracy with which waveform peak intensity and spatiotemporal peak location can be detected using multi- or hyperspectral lidar instrumentation. Different full-waveform reconstruction algorithms with varying characteristics are implemented on simulated Gaussian laser pulse data sets obtained with varying sampling frequencies ranging between 1 GHz and 5 GHz. The data sets are analyzed, and an optimized digitization frequency is found based on observation of algorithmic parameter retrieval accuracy. The accuracy of the full-waveform retrieval in relation to the type of algorithm used and its robustness related to the digitization frequency are also discussed. We find that optimizing the algorithmic processing of multi-wavelength sampled radiance data recorded by multispectral and hyperspectral lidar instruments significantly improves the accuracy of reflectance retrieval and target material characterization capabilities of the instrument.