A Classification Method for Airborne Full-Waveform LiDAR Systems Based on a Gramian Angular Field and Convolution Neural Networks

The data processing of airborne full-waveform light detection and ranging (LiDAR) systems has become a research hotspot in the LiDAR field recent years. However, accuracy reliability classification remain challenge. manual features deep learning techniques existing methods cannot fully utilize temporal spatial information full waveform. On premise preserving dependencies, we convert them into Gramian angular summation (GASF) images using polar coordinate method. By introducing attention modules neural network, emphasize importance location texture GASF images. Finally, use open source simulated to evaluate impact different network architectures transformation methods. Compared with performance state-of-art method, our proposed method can achieve higher precision F1 scores. results suggest that transforming waveform module outperformed other