Vehicle Detection from 3D Lidar Using Fully Convolutional Network

Abstract

Convolutional network techniques have recently achieved great success in vision based detection tasks. This paper introduces the recent development of our research on transplanting the fully convolutional network technique to the detection tasks on 3D range scan data. Specifically, the scenario is set as the vehicle detection task from the range data of Velodyne 64E lidar. We proposes to present the data in a 2D point map and use a single 2D end-to-end fully convolutional network to predict the objectness confidence and the bounding boxes simultaneously. By carefully design the bounding box encoding, it is able to predict full 3D bounding boxes even using a 2D convolutional network. Experiments on the KITTI dataset shows the state-of-the-art performance of the proposed method.

Robotics: Science and Systems, 2016

References in corpus (2)

• Fully Convolutional Networks for Semantic Segmentation
• Learning Rich Features from RGB-D Images for Object Detection and Segmentation

Cited by in corpus (73)

• 3D Object Detection for Autonomous Driving: A Survey
• VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection
• Joint Activity Recognition and Indoor Localization with WiFi Fingerprints
• SegMap: 3D Segment Mapping using Data-Driven Descriptors
• Joint 3D Proposal Generation and Object Detection from View Aggregation
• End-to-End Multi-View Fusion for 3D Object Detection in LiDAR Point Clouds
• 3D Object Detection Using Scale Invariant and Feature Reweighting Networks
• Towards Robust LiDAR-based Perception in Autonomous Driving: General Black-box Adversarial Sensor Attack and Countermeasures
• Orthographic Feature Transform for Monocular 3D Object Detection
• Complex-YOLO: Real-time 3D Object Detection on Point Clouds
• A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation
• Multi-View 3D Object Detection Network for Autonomous Driving
• OREOS: Oriented Recognition of 3D Point Clouds in Outdoor Scenarios
• SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud
• PointFusion: Deep Sensor Fusion for 3D Bounding Box Estimation
• RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving
• Spatially-Aware Graph Neural Networks for Relational Behavior Forecasting from Sensor Data
• Deep Learning for 3D Point Clouds: A Survey
• Roadside Lidar Vehicle Detection and Tracking Using Range And Intensity Background Subtraction
• Radar-Camera Sensor Fusion for Joint Object Detection and Distance Estimation in Autonomous Vehicles
• TANet: Robust 3D Object Detection from Point Clouds with Triple Attention
• SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation
• Deep Learning for LiDAR Point Clouds in Autonomous Driving: A Review
• 3D Fully Convolutional Network for Vehicle Detection in Point Cloud
• A General Pipeline for 3D Detection of Vehicles
• CAE-LO: LiDAR Odometry Leveraging Fully Unsupervised Convolutional Auto-Encoder for Interest Point Detection and Feature Description
• A survey of Object Classification and Detection based on 2D/3D data
• RST-MODNet: Real-time Spatio-temporal Moving Object Detection for Autonomous Driving
• Fusing Bird View LIDAR Point Cloud and Front View Camera Image for Deep Object Detection
• Triangulation Learning Network: from Monocular to Stereo 3D Object Detection
• An LSTM Approach to Temporal 3D Object Detection in LiDAR Point Clouds
• Training a Fast Object Detector for LiDAR Range Images Using Labeled Data from Sensors with Higher Resolution
• Improving 3D Object Detection with Channel-wise Transformer
• Efficient Deep Neural Networks
• milliEgo: Single-chip mmWave Radar Aided Egomotion Estimation via Deep Sensor Fusion
• LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR
• LATTE: Accelerating LiDAR Point Cloud Annotation via Sensor Fusion, One-Click Annotation, and Tracking
• Feature Sharing and Integration for Cooperative Cognition and Perception with Volumetric Sensors
• A Spike Learning System for Event-driven Object Recognition
• Object Detection based on LIDAR Temporal Pulses using Spiking Neural Networks
• 3D-DEEP: 3-Dimensional Deep-learning based on elevation patterns forroad scene interpretation
• An Overview Of 3D Object Detection
• Real-time 3D object proposal generation and classification under limited processing resources
• Unsupervised Data Imputation via Variational Inference of Deep Subspaces
• Class-specific Anchoring Proposal for 3D Object Recognition in LIDAR and RGB Images
• Dynamic Edge Weights in Graph Neural Networks for 3D Object Detection
• 3D Object Detection and High-Resolution Traffic Parameters Extraction Using Low-Resolution LiDAR Data
• Architecting Dependable Learning-enabled Autonomous Systems: A Survey
• Are We Hungry for 3D LiDAR Data for Semantic Segmentation? A Survey and Experimental Study
• Local Aggressive Adversarial Attacks on 3D Point Cloud
• Learning Orientation-Estimation Convolutional Neural Network for Building Detection in Optical Remote Sensing Image
• SurfConv: Bridging 3D and 2D Convolution for RGBD Images
• DeepPCO: End-to-End Point Cloud Odometry through Deep Parallel Neural Network
• Cooperative LIDAR Object Detection via Feature Sharing in Deep Networks
• LiDAR R-CNN: An Efficient and Universal 3D Object Detector
• Unsupervised Object Detection with LiDAR Clues
• Cirrus: A Long-range Bi-pattern LiDAR Dataset
• VR3Dense: Voxel Representation Learning for 3D Object Detection and Monocular Dense Depth Reconstruction
• RUHSNet: 3D Object Detection Using Lidar Data in Real Time
• Safety-Aware Hardening of 3D Object Detection Neural Network Systems
• S3Net: 3D LiDAR Sparse Semantic Segmentation Network
• Joint Spatial-Temporal Optimization for Stereo 3D Object Tracking
• Novel Perception Algorithmic Framework For Object Identification and Tracking In Autonomous Navigation
• Deep Lidar CNN to Understand the Dynamics of Moving Vehicles
• Range Adaptation for 3D Object Detection in LiDAR
• All-Weather Object Recognition Using Radar and Infrared Sensing
• 3D Lidar Mapping Relative Accuracy Automatic Evaluation Algorithm
• 3D Objectness Estimation via Bottom-up Regret Grouping
• Non-parametric Memory for Spatio-Temporal Segmentation of Construction Zones for Self-Driving
• Know Your Surroundings: Panoramic Multi-Object Tracking by Multimodality Collaboration
• To the Point: Efficient 3D Object Detection in the Range Image with Graph Convolution Kernels
• Empowering cyberphysical systems of systems with intelligence
• Depth-Adapted CNN for RGB-D cameras