For full-waveform LiDAR echo signals, a high efficient and accurate decomposition method based on a dense (Full-waveform Dense Connection Network, FDCN) and a residual neural networks (Full-waveform Deep Residual Network, FDRN) is proposed in this paper.

Instead of using multiple sets of measurements, we discuss a CNN with one set of data to obtain temporal super-resolution in full-waveform LiDAR. The super-resolution results can enhance further waveform decomposition or classification performance.

No abstract

Different from conventional decomposition methods which utilize several steps to obtain the final result, a selfattention based neural network, Attention Full-waveform Decomposition Network (AFD-Net), is discussed in this paper for endto-end full-waveform LiDAR signal decomposition. In existing LiDAR waveform decomposition methods, complicate functional models are used to fit echo components. Thus, the echo decomposition problem can be translated into a function approximation task. Recent studies present great progress in estimating the parameters of fitting models, hence in the final decomposition results. However, the shape of received echo components are always irregular. None of the parametric functional models can fit the received echo components perfectly, which leads to unavoidable errors in the initial step of echo decomposition. In this paper, we propose an end-to-end net work AFD-Net to solve the echo decomposition problem without assuming any parametric functional models. AFD-Net is consisted with two modules, the classification module and the decomposition module. The former module is used to determine the number of echo components in a received waveform. Then the decomposition module is used to output the echo components. By experiments, we have a classification accuracy 96% using the first module. The average R 2 coefficient for the decomposed echo components is 0.9799. In addition, there are no public datasets for the waveform decomposition task available. Thus, another contribution of our work is to develop a tool to generate synthetic full-waveform LiDAR signals, which can help researchers to construct their own dataset for related works. All of our source codes are available in the following site: https://github.com/ZedFm/AFD-Net

For full-waveform (FW) LiDAR signals, conventional echo decomposition methods use complicated filtering or de-noising algorithms for signal pre-processing. However, the speed and accuracy of these algorithms are limited. In this paper, we study a highly efficient and accurate decomposition method based on the FW dense connection network (FDCN) or FW deep residual network (FDRN). FDCN is a lightweight and efficient network for SNR higher than 24 dB, while FDRN is a deeper neural network with multiple residual blocks and works well for low SNR such as 12 dB. We compare FDCN and FDRN with other conventional methods. With FDCN and FDRN, the mean error for estimating an echo peak location is under 0.2 ns, while the amplitude error is under 5 mV when the dynamic range is 0 ∼ 100 m V . Both errors are much lower than the values using conventional methods.