On the segmentation of 3D LIDAR point clouds

By B. Douillard, James Patrick Underwood, Noah Kuntz, Vsevolod Vlaskine, Alastair James Quadros, P. Morton, Alon Frenkel. 2011

499 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Environmental Engineering, Geology

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. Bo Li
2. Roland Yves Siegwart
3. Xiangyu Yue
4. Kurt Keutzer
5. Bichen Wu
6. Ingmar Posner
7. Senthil Kumar Yogamani
8. Jens Behley

In Environmental Engineering

1. Bo Li
2. Tian Xia
3. Ingmar Posner
4. Dominic Zeng Wang
5. Bichen Wu
6. Kurt Keutzer
7. Xiangyu Yue
8. Tianlei Zhang

In Geology

1. Bo Li
2. Tian Xia
3. Tianlei Zhang
4. Jens Behley
5. Ingmar Posner
6. Volker Steinhage
7. Armin Bernd Cremers
8. Dominic Zeng Wang

Citations in the corpus

• 3D fully convolutional network for vehicle detection in point cloud (2016, background): “For example, [8], [14], [25] remove the ground plane and cluster the remained segments as potential object proposals.”
• Vehicle Detection from 3D Lidar Using Fully Convolutional Network (2016, methodology): “In some early works, rule-based segmentation is suggested for specific scene [10, 20, 5].”
• SegMatch: Segment based place recognition in 3D point clouds (2016, methodology): “The filtered point cloud is finally segmented into a set of point clusters Ci using the "Cluster-All Method" of [22].”
• Fast segmentation of 3D point clouds: A paradigm on LiDAR data for autonomous vehicle applications (2017, background): “[5] are introducing GP-INSAC, a gaussian-process based iterative scheme that classifies points in ground and non-ground based on the variance of their height from the mean of a gaussian distribution.”
• RGB and LiDAR fusion based 3D Semantic Segmentation for Autonomous Driving (2019, background): “Point cloud segmentation has been studied in [4][10][23].”
• What could move? Finding cars, pedestrians and bicyclists in 3D laser data (2012, background): “[11], where the existence of a ground plane is assumed and object segments are derived in an unsupervised fashion from non-ground data only.”
• All Weather Perception: Joint Data Association, Tracking, and Classification for Autonomous Ground Vehicles (2016, methodology): “Lidar point cloud processing methods have improved to enable accurate segmentation and classification of high resolution 3d point clouds in real-time (Douillard et al., 2011) and (Korchev et al., 2013).”
• SqueezeSegV2: Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloud (2018, methodology): “Non-deep-learning methods [1], [7], [8] usually involve several stages such as ground removal, clustering, and classification.”
• Deep Learning Advances in Computer Vision with 3D Data (2017, methodology): “Furthermore, clustering algorithms [Douillard et al. 2011], 3D Hough Trans-form [Borrmann et al. 2011], RANSAC [Schnabel et al. 2007], and probabilistic models, such as Markov Random Fields (MRFs) and Conditional Random Fields (CRFs), have also been employed (e.g., Valentin et al. [2013]).”
• SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud (2017, methodology): “[2] summarized several approaches based on iterative algorithms such as RANSAC and GP-INSAC for ground removal.”
• Laser-based segment classification using a mixture of bag-of-words (2013, background): “All approaches share a non-trivial selection of suitable parameters, which is usually specific to the task and object classes of interest [15].”