Real-time object detection using a sparse 4-layer LIDAR

Muresan, Mircea Paul; Nedevschi, Sergiu; Giosan, Ion

doi:10.1109/iccp.2017.8117023

Cited by 15 publications

(15 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To tackle the sensor-fusion computational problem, this [22] proposed an earlyfusion method to fuse both camera and LiDAR with only one backbone, attaining a good balance between accuracy and efficiency. Other methods such as [23] solve the problem of data correction and temporal point cloud fusion for object detection using only 4-layer LiDAR. However, the discussed state-of-the-art approaches' performance depends on the large-scale training dataset and ground truth labels.…”

Section: D Object Detectionmentioning

confidence: 99%

Transfer Learning Based Semantic Segmentation for 3D Object Detection from Point Cloud

Imad

Doukhi

Lee

2021

Sensors

View full text Add to dashboard Cite

Three-dimensional object detection utilizing LiDAR point cloud data is an indispensable part of autonomous driving perception systems. Point cloud-based 3D object detection has been a better replacement for higher accuracy than cameras during nighttime. However, most LiDAR-based 3D object methods work in a supervised manner, which means their state-of-the-art performance relies heavily on a large-scale and well-labeled dataset, while these annotated datasets could be expensive to obtain and only accessible in the limited scenario. Transfer learning is a promising approach to reduce the large-scale training datasets requirement, but existing transfer learning object detectors are primarily for 2D object detection rather than 3D. In this work, we utilize the 3D point cloud data more effectively by representing the birds-eye-view (BEV) scene and propose a transfer learning based point cloud semantic segmentation for 3D object detection. The proposed model minimizes the need for large-scale training datasets and consequently reduces the training time. First, a preprocessing stage filters the raw point cloud data to a BEV map within a specific field of view. Second, the transfer learning stage uses knowledge from the previously learned classification task (with more data for training) and generalizes the semantic segmentation-based 2D object detection task. Finally, 2D detection results from the BEV image have been back-projected into 3D in the postprocessing stage. We verify results on two datasets: the KITTI 3D object detection dataset and the Ouster LiDAR-64 dataset, thus demonstrating that the proposed method is highly competitive in terms of mean average precision (mAP up to 70%) while still running at more than 30 frames per second (FPS).

show abstract

Section: D Object Detectionmentioning

confidence: 99%

Transfer Learning Based Semantic Segmentation for 3D Object Detection from Point Cloud

Imad

Doukhi

Lee

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…In the presented pipeline, the modules Object Spatial and Temporal Alignment and LIDAR Motion Correction deal with the spatio-temporal alignment of raw sensor data to a reference timestamp, given by the front fisheye camera. This alignment is performed at object level for the long-range RADAR and trifocal camera objects, as described in Section 3.2., and at point cloud level in the LIDAR motion correction module [46].…”

Section: General System Overviewmentioning

confidence: 99%

“…In the presented pipeline, the modules Object Spatial and Temporal Alignment and LIDAR Motion Correction deal with the spatiotemporal alignment of raw sensor data to a reference timestamp, given by the front fisheye camera. This alignment is performed at object level for the long-range RADAR and trifocal camera objects, as described in Section 3.2., and at point cloud level in the LIDAR motion correction module [46]. The motion-corrected point clouds are projected, using the Points Projection module, onto the intensity image and onto a semantic segmentation image obtained as described in [47] and given by the Semantic Segmentation module, to obtain an enhanced point cloud where each 3D point will contain semantic information as well as color information.…”

Section: General System Overviewmentioning

confidence: 99%

Stabilization and Validation of 3D Object Position Using Multimodal Sensor Fusion and Semantic Segmentation

Muresan

Giosan

Nedevschi

2020

Sensors

Self Cite

View full text Add to dashboard Cite

The stabilization and validation process of the measured position of objects is an important step for high-level perception functions and for the correct processing of sensory data. The goal of this process is to detect and handle inconsistencies between different sensor measurements, which result from the perception system. The aggregation of the detections from different sensors consists in the combination of the sensorial data in one common reference frame for each identified object, leading to the creation of a super-sensor. The result of the data aggregation may end up with errors such as false detections, misplaced object cuboids or an incorrect number of objects in the scene. The stabilization and validation process is focused on mitigating these problems. The current paper proposes four contributions for solving the stabilization and validation task, for autonomous vehicles, using the following sensors: trifocal camera, fisheye camera, long-range RADAR (Radio detection and ranging), and 4-layer and 16-layer LIDARs (Light Detection and Ranging). We propose two original data association methods used in the sensor fusion and tracking processes. The first data association algorithm is created for tracking LIDAR objects and combines multiple appearance and motion features in order to exploit the available information for road objects. The second novel data association algorithm is designed for trifocal camera objects and has the objective of finding measurement correspondences to sensor fused objects such that the super-sensor data are enriched by adding the semantic class information. The implemented trifocal object association solution uses a novel polar association scheme combined with a decision tree to find the best hypothesis–measurement correlations. Another contribution we propose for stabilizing object position and unpredictable behavior of road objects, provided by multiple types of complementary sensors, is the use of a fusion approach based on the Unscented Kalman Filter and a single-layer perceptron. The last novel contribution is related to the validation of the 3D object position, which is solved using a fuzzy logic technique combined with a semantic segmentation image. The proposed algorithms have a real-time performance, achieving a cumulative running time of 90 ms, and have been evaluated using ground truth data extracted from a high-precision GPS (global positioning system) with 2 cm accuracy, obtaining an average error of 0.8 m.

show abstract

“…It is important to achieve robust and accurate scene understanding for video surveillance and autonomous driving [1]. Camera-based [2,3], LiDAR-based [4], radar-based [5] and sensor fusion-based [6] methods are widely used in scene perception and semantic information recognition of moving targets. Although a visual camera cannot obtain the depth information of targets and the environment, it is unburdened by the cost of the equipment or the complicated sensor fusion strategy.…”

Section: Introductionmentioning

confidence: 99%

Trajectory dimensionality reduction and hyperparameter settings of DBSCAN for trajectory clustering

Long

et al. 2022

IET Intelligent Trans Sys

View full text Add to dashboard Cite

The density-based spatial clustering of application with noise (DBSCAN) algorithm has good robustness and is widely employed to cluster vehicle trajectories for vehicle movement pattern recognition. However, the distance or similarity between two trajectories varies from tens to hundreds of thousands, and there is no effective method for determining the values of the hyperparameters eps and MinPts of DBSCAN. In addition, with increasing sizes of trajectory datasets, some trajectory clustering methods that directly analyse points and line segments incur large computational costs and time overhead. To solve these two dilemmas, the authors propose an effective trajectory dimensionality reduction method and a DBSCAN hyperparameter initial value setting method. The trajectory dimensionality reduction algorithm processes trajectories with different lengths into the same dimensionality (the same number of feature points). The reserved points preserve the spatial and temporal information of these trajectories as much as possible. The DBSCAN hyperparameter initial value setting algorithm obtains the effective initial values of eps and MinPts for facilitating subsequent adjustments. Finally, we validate these proposed methods on two trajectory datasets collected from two real-world scenes, and the experimental results are promising and effective.

show abstract

Real-time object detection using a sparse 4-layer LIDAR

Cited by 15 publications

References 11 publications

Transfer Learning Based Semantic Segmentation for 3D Object Detection from Point Cloud

Transfer Learning Based Semantic Segmentation for 3D Object Detection from Point Cloud

Stabilization and Validation of 3D Object Position Using Multimodal Sensor Fusion and Semantic Segmentation

Trajectory dimensionality reduction and hyperparameter settings of DBSCAN for trajectory clustering

Contact Info

Product

Resources

About