Multi-Target reflectance reconstruction based on waveform decomposition with Full-Waveform hyperspectral liDAR

Full-waveform hyperspectral LiDAR (FWHSL) can simultaneously obtain three-dimensional structural and continuous spectral information of targets, making it a promising technique for quantitative target characterization. To fully realize this potential, accurate reflectance extraction is a crucial prerequisite for subsequent applications such as material identification, target classification, and biochemical parameter retrieval. However, in practical FWHSL measurements, a single laser footprint may simultaneously illuminate multiple adjacent targets. The returned waveform then contains contributions from different sub-targets, with their spectral responses and spatial positions coupled in the same echo signal. This coupling makes conventional reflectance extraction methods, which are usually designed for single-target echoes, prone to errors and may result in biased reflectance reconstruction. First, a reference set of single-target reflectance curves is established. Second, a mixture model of skew-normal distributions combined with the trust region optimization algorithm is used to decompose the overlapping echoes and extract the geometric and intensity parameters of each sub-target. Finally, by combining the LiDAR equation with the Gaussian laser spot energy distribution model, a correction factor for partially illuminated energy is derived, and the reflectance of each sub-target is reconstructed using the least-squares method. To validate the method, two types of experiments (simulation and measurement experiments) were conducted. Simulation experiments were conducted to generate dual-target overlapping echoes with neighbor distances (NDs) of 5–80 cm using the USGS spectral library. Measurement experiments were then carried out with the FWHSL system on two target combinations, with the same ND (5–80 cm) at incidence angles of 0 ∘ , 30 ∘ , and 60 ∘ . Results indicate that when the neighbor distance exceeds 13 cm, the method can reconstruct sub-target reflectance with high accuracy.