LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving

Abstract

The capability to detect objects is a core part of autonomous driving. Due to sensor noise and incomplete data, perfectly detecting and localizing every object is infeasible. Therefore, it is important for a detector to provide the amount of uncertainty in each prediction. Providing the autonomous system with reliable uncertainties enables the vehicle to react differently based on the level of uncertainty. Previous work has estimated the uncertainty in a detection by predicting a probability distribution over object bounding boxes. In this work, we propose a method to improve the ability to learn the probability distribution by considering the potential noise in the ground-truth labeled data. Our proposed approach improves not only the accuracy of the learned distribution but also the object detection performance.

Sensor Fusion for Joint 3D Object Detection and Semantic Segmentation

“…Typically, previous work [11,15,31,32,34,35] discretizes the LiDAR points into 3D voxels and performs convolutions in the bird's eye view (BEV). Only a few methods [14,18,29] utilize the native range view (RV) of the LiDAR sensor. In terms of 3D object detection, BEV methods have traditionally achieved higher performance than RV methods.…”

“…On the other hand, RV methods are usually more computationally efficient because the RV is a compact representation of the LiDAR data where the BEV is sparse. Recently, [18] demonstrated that a RV method can be both efficient and obtain state-of-the-art performance when trained on a significantly large dataset. Furthermore, they showed that a RV detector can produce more accurate detections on small objects, such as pedestrians and bikes.…”

“…In this paper, we propose an efficient method for fusing 2D image data and 3D LiDAR data, and we leverage this approach to improve LaserNet, an existing state-of-the-art Li-DAR based 3D object detector [18]. Our sensor fusion technique is efficient allowing us to maintain LaserNet's low runtime.…”

Training a Fast Object Detector for LiDAR Range Images Using Labeled Data from Sensors with Higher Resolution

“…To ensure invariance under rotation around the sensor origin, the box position and object orientation are represented in a coordinate frame that is aligned with the direction of sight to the corresponding point. The resulting box representation is effectively a three-dimensional extension of the representation in [5]. This representation significantly reduces the number of output channel compared to the separate prediction of all corner points as found in [14].…”

“…This is somewhat expected, as this layer may not only remove wrongly predicted objects from pixels around the edges of objects in the range image but also reduce the objectness score of distant, small, or partially visible objects. These might be handled better by approaches similar to the mean shift clustering used in [5] (at the cost of complexity and runtime). If no orientation anchors are used and only a single set of regression channels is predicted, the detection performance drops significantly.…”