Automatic Labeling to Generate Training Data for Online LiDAR-Based Moving Object Segmentation

Understanding the scene is key for autonomously navigating vehicles, and the ability to segment the surroundings online into moving and non-moving objects is a central ingredient of this task. Often, deep learning-based methods are used to perform moving object segmentation (MOS). The performance of these networks, however, strongly depends on the diversity and amount of
labeled
training data—information that may be costly to obtain. In this letter, we propose an automatic data labeling pipeline for 3D LiDAR data to save the extensive manual labeling effort and to improve the performance of existing learning-based MOS systems by automatically annotation training data. Our proposed approach achieves this by processing the data offline in batches, i.e., it is not designed to run online on a vehicle. It labels the actually moving objects such as driving cars and pedestrians as moving. In contrast, the non-moving objects, e.g., parked cars, lamps, roads, or buildings, are labeled as static. We show that this approach allows us to label LiDAR data highly effectively and compare our results to those of other label generation methods. We also train a deep neural network with our automatically generated labels and achieve comparable performance to the one trained with manual labels on the same data—and an even better performance when using additional datasets with labels generated by our approach. Furthermore, we evaluate our method on multiple datasets using different sensors, and our experiments indicate that our method can generate labels in different outdoor environments.