Object classification in radar using ensemble methods

Lombacher, Jakob; Hahn, Markus; Dickmann, Jürgen; Wöhler, Christian

doi:10.1109/icmim.2017.7918863

Cited by 36 publications

(9 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Even though CNNs function extraordinarily well on images, they can also be tried and applied to other sensors that can yield image-like data [ 108 ]. The two-dimensional radar grid representations accumulated according to different occupancy grid map algorithms have already been exploited in deep learning domains for various autonomous system tasks, such as static object classification [ 109 , 110 , 111 , 112 , 113 , 114 ] and dynamic object classification [ 115 , 116 , 117 ]. In this case, the objects denote any road user within an autonomous system environment, like the pedestrian, vehicles, motorcyclists, etc.…”

Section: Detection and Classification Of Radar Signals Using Deep mentioning

confidence: 99%

“…The candidate’s region was extracted, and two random forest classifiers were trained to confirm the parked vehicle’s presence. Subsequently, Lambacher et al [ 110 , 111 ] presented a classification technique for static object recognition based on radar signals and DCNNs. The occupancy grid algorithm was used to accumulate the radar data into grid representations.…”

Section: Detection and Classification Of Radar Signals Using Deep mentioning

confidence: 99%

See 1 more Smart Citation

Application of Deep Learning on Millimeter-Wave Radar Signals: A Review

Abdu

Zhang

et al. 2021

Sensors

View full text Add to dashboard Cite

The progress brought by the deep learning technology over the last decade has inspired many research domains, such as radar signal processing, speech and audio recognition, etc., to apply it to their respective problems. Most of the prominent deep learning models exploit data representations acquired with either Lidar or camera sensors, leaving automotive radars rarely used. This is despite the vital potential of radars in adverse weather conditions, as well as their ability to simultaneously measure an object’s range and radial velocity seamlessly. As radar signals have not been exploited very much so far, there is a lack of available benchmark data. However, recently, there has been a lot of interest in applying radar data as input to various deep learning algorithms, as more datasets are being provided. To this end, this paper presents a survey of various deep learning approaches processing radar signals to accomplish some significant tasks in an autonomous driving application, such as detection and classification. We have itemized the review based on different radar signal representations, as it is one of the critical aspects while using radar data with deep learning models. Furthermore, we give an extensive review of the recent deep learning-based multi-sensor fusion models exploiting radar signals and camera images for object detection tasks. We then provide a summary of the available datasets containing radar data. Finally, we discuss the gaps and important innovations in the reviewed papers and highlight some possible future research prospects.

show abstract

Section: Detection and Classification Of Radar Signals Using Deep mentioning

confidence: 99%

Section: Detection and Classification Of Radar Signals Using Deep mentioning

confidence: 99%

Application of Deep Learning on Millimeter-Wave Radar Signals: A Review

Abdu

Zhang

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…To reduce the sparsity of radar point clouds, a common approach involves the integration of radar reflections over time, so to generate occupancy or amplitude grid-maps [24]. These maps consist in structured, image-like inputs-whose cells contain information about occupancy or measured RCS-, thus enabling the use of standard CNNs [16,17]. However, despite this approach is well-suited for static-objects, it fails on dynamic ones, as the object velocity needs to be tracked to avoid generating a tail of reflections.…”

Section: Deep Learning On Radar Datamentioning

confidence: 99%

A Neural Network Based System for Efficient Semantic Segmentation of Radar Point Clouds

Cennamo

Kaestner

Kummert

2021

Neural Process Lett

View full text Add to dashboard Cite

The last decade has witnessed important advancements in the field of computer vision and scene understanding, enabling applications such us autonomous vehicles. Radar is a commonly adopted sensor in automotive industry, but its suitability to machine learning techniques still remains an open question. In this work, we propose a neural network (NN) based solution to efficiently process radar data. We introduce RadarPCNN, an architecture specifically designed for performing semantic segmentation on radar point clouds. It uses PointNet$$++$$ + + as a building-block—enhancing the sampling stage with mean-shift—and an attention mechanism to fuse information. Additionally, we propose a machine learning radar pre-processing module that confers the network the ability to learn from radar features. We show that our solutions are effective, yielding superior performance than the state-of-the-art.

show abstract

“…Areas of interest within the SGs were extracted with a clustering algorithm and then investigated with analytical methods. On the other hand, [7] uses convolutional neural networks (CNN) to investigate the image details. Finally, [4] and [8] achieved good results with complete pixel-wise classifications using extensive CNNs, thus avoiding the error-prone clustering step.…”

Section: Introductionmentioning

confidence: 99%

Semantic Segmentation on 3D Occupancy Grids for Automotive Radar

et al. 2020

View full text Add to dashboard Cite

Radar sensors have great advantages over other sensors in estimating the motion states of moving objects, because they detect velocity components within one measurement cycle. Moreover, numerous successes have already been achieved regarding the classification of such objects. However, the advantage of instantaneous velocity measurement is lost when detecting static objects, so that their classification is much more demanding. In this paper, we use semantic segmentation networks to distinguish between frequently occurring infrastructure objects. The resulting semantic grids provide a location-based classification of the vehicle environment. Since even modern radars have a significantly poorer angular resolution than lidars, the relatively thin radar point cloud is accumulated in advance and transformed into 2D or 3D grids that act as network inputs. Occupancy grids are particularly advantageous here, since they calculate not only the obstacles but also the free spaces. With suitable parameter selection, which is very challenging due to the complexity of radar measurement, the resulting grids allow for good association with camera images. Finally, in order to evaluate possible advantages of 3D grids as network input with respect to the segmentation result, we created and evaluated a simulation dataset and two different real-world datasets in car parks and on motorways. As a result, Jaccard coefficients between 81% and 88% were achieved, depending on the dataset. It was also found that 3D input images lead to improvements in the car park dataset.

show abstract

Object classification in radar using ensemble methods

Cited by 36 publications

References 10 publications

Application of Deep Learning on Millimeter-Wave Radar Signals: A Review

Application of Deep Learning on Millimeter-Wave Radar Signals: A Review

A Neural Network Based System for Efficient Semantic Segmentation of Radar Point Clouds

Semantic Segmentation on 3D Occupancy Grids for Automotive Radar

Contact Info

Product

Resources

About