Neural LiDAR Fields for Novel View Synthesis
This work addresses the domain gap in LiDAR data synthesis for applications like autonomous driving, though it is an incremental improvement over existing neural field methods.
The paper tackles the problem of synthesizing realistic LiDAR scans from novel viewpoints by proposing Neural Fields for LiDAR (NFL), which combines neural fields with a detailed LiDAR sensor model. The results show that NFL outperforms existing methods on LiDAR novel view synthesis and improves downstream tasks like registration and semantic segmentation.
We present Neural Fields for LiDAR (NFL), a method to optimise a neural field scene representation from LiDAR measurements, with the goal of synthesizing realistic LiDAR scans from novel viewpoints. NFL combines the rendering power of neural fields with a detailed, physically motivated model of the LiDAR sensing process, thus enabling it to accurately reproduce key sensor behaviors like beam divergence, secondary returns, and ray dropping. We evaluate NFL on synthetic and real LiDAR scans and show that it outperforms explicit reconstruct-then-simulate methods as well as other NeRF-style methods on LiDAR novel view synthesis task. Moreover, we show that the improved realism of the synthesized views narrows the domain gap to real scans and translates to better registration and semantic segmentation performance.