RangeNet ++: Fast and Accurate LiDAR Semantic Segmentation

By Andres Milioto, Ignacio Vizzo, Jens Behley, Cyrill Stachniss. IEEE/RJS International Conference on Intelligent RObots and Systems, 2019

1,313 citations recorded by the source corpus.

Perception in autonomous vehicles is often carried out through a suite of different sensing modalities. Given the massive amount of openly available labeled RGB data and the advent of high-quality deep learning algorithms for image-based recognition, high-level semantic perception tasks are pre-dominantly solved using high-resolution cameras. As a result of that, other sensor modalities potentially useful for this task are often ignored. In this paper, we push the state of the art in LiDAR-only semantic segmentation forward in order to provide another independent source of semantic information to the vehicle. Our approach can accurately perform full semantic segmentation of LiDAR point clouds at sensor frame rate. We exploit range images as an intermediate representation in combination with a Convolutional Neural Network (CNN) exploiting the rotating LiDAR sensor model. To obtain accurate results, we propose a novel post-processing algorithm that deals with problems arising from this intermediate representation such as discretization errors and blurry CNN outputs. We implemented and thoroughly evaluated our approach including several comparisons to the state of the art. Our experiments show that our approach outperforms state-of-the-art approaches, while still running online on a single embedded GPU. The code can be accessed at https://github.com/PRBonn/lidar-bonnetal.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Instrumentation

View Paper: https://www.semanticscholar.org/paper/64e23f006fda19bbbbc324b1353851ebca95aa02

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Yulan Guo
2. Cyrill Stachniss
3. Philippe Giguère
4. Andres Milioto Emanuele Palazzolo
5. Xieyuanli Chen
6. Jens Behley
7. J. Ibañez-Guzmán
8. You Li

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Kurt Keutzer
3. Dahua Lin
4. Hongsheng Li
5. Dacheng Tao
6. Xiaogang Wang
7. Bichen Wu
8. Ming-Hsuan Yang

In Instrumentation

1. J. Ibañez-Guzmán
2. You Li
3. Wei Zhan
4. Zining Wang
5. Masayoshi Tomizuka
6. Cyrill Stachniss
7. Jens Behley
8. Hongsheng Li

Citations in the corpus

• PandaSet: Advanced Sensor Suite Dataset for Autonomous Driving (2021, methodology): “Since PandaSet also provides ground truth labels for LiDAR point cloud segmentation, we establish its baseline using RangeNet53, using only the network, without the post-processing [24].”
• RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation (2021, methodology): “For instance, RangeNet++ [23] draws on the DarkNet backbone in YOLOv3 [27] as the extractor and is post-processed by using an accelerated KNN. SalsaNext [12] takes Sal- sanet [1] as baseline and presents an uncertainty-aware mechanism for point feature learning.”
• 2DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR Point Clouds (2022, methodology): “PointPainting [15] exploits the segmentation logits of images and projects them to the LiDAR space by bird’s-eye projection [23] or spherical projection [45] for LiDAR network performance improvement.”
• Deep Learning for 3D Point Clouds: A Survey (2019, methodology): “methods in Fig.8. 4.1.1 Projection-based Networks Intermediate regular representations can be organised or classfied into multi-view representation [148], [149], spherical representation [150], [151], [152], volumetric representation [153], [154], [155], permutohedral lattice representation [156], [1”
• PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation (2020, methodology): “For spherical projection, we followed the setup of projecting point clouds with zenith angles ranging from −25◦ to 3◦ into [64, 2048] grids in the projected sphere plane as in [19].”
• SuMa++: Efficient LiDAR-based Semantic SLAM (2019, methodology): “The point-wise labels are provided by RangeNet++ using spherical projections of the point clouds.”
• Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentation (2020, methodology): “RangeNet++ [27] 65.”
• (AF)2-S3Net: Attentive Feature Fusion with Adaptive Feature Selection for Sparse Semantic Segmentation Network (2021, methodology): “In addition to a novel CNN, RangeNet++ [21] proposes an efficient way of predicting labels for the full point cloud using a fast implementation of K-nearest neighbour (KNN).”
• SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation (2020, methodology): “Previous methods based on spherical projection [56,58,30] treat projected LiDAR images as RGB images and process them with standard convolution as”
• Deep Multi-Modal Object Detection and Semantic Segmentation for Autonomous Driving: Datasets, Methods, and Challenges (2019, neutral): “[51]–[56]).”
• RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (2019, methodology): “2 RangeNet53++ [40] 52.”
• Lidar for Autonomous Driving: The Principles, Challenges, and Trends for Automotive Lidar and Perception Systems (2020, methodology): “Based on this dataset, RangeNet [51] demonstrated fascinating performance and speed from a not complicated DNN structure.”