Gaussian decomposition method for full waveform data of LiDAR base on neural network
Abstract The full waveform data from airborne LiDAR (Light Detection and Ranging) provides information on the distance of the target. Accurately extracting the ranging information from the full waveform data is crucial for generating point clouds. This paper introduces a method for Gaussian decompos...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-02-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-024-82543-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract The full waveform data from airborne LiDAR (Light Detection and Ranging) provides information on the distance of the target. Accurately extracting the ranging information from the full waveform data is crucial for generating point clouds. This paper introduces a method for Gaussian decomposition of full waveform data using a convolutional neural network. The method employs an improved densely connected convolutional neural network and the EM (Expectation Maximization) algorithm to extract information from the data. The method involves two key steps. First, The FWDN network preprocesses the full waveform data to enhance signal quality by reducing noise, and then the improved EM algorithm extracts Gaussian parameters (amplitude, expectation, and full width at half maximum) to obtain ranging information. Based on simulation and measured data, the decomposition success rate of this method is more than 98% with an average range accuracy of less than 1.5 cm compared to other methods. The method has significant potential for application in the field of mapping, 3D modeling. |
|---|---|
| ISSN: | 2045-2322 |