Semantic Segmentation of Outdoor Areas Using 3D Moment Invariants and Contextual Cues

Sickert, Sven; Denzler, Joachim

doi:10.1007/978-3-319-66709-6_14

Cited by 3 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Point-wise approaches are computationally expensive but can offer high details in scenarios with low point densities. The authors of [17] propose to compute 3D moment invariant features for each occurring point. They are invariant under scaling, rotation, and translation.…”

Section: B Points-wise Approachesmentioning

confidence: 99%

“…It consists of 467, 211 points recorded using a terrestrial LiDAR scanner. The dataset covers an area containing multiple trees with labels [34] labeled with four classes using 2-fold cross-validation based on the split proposed in [17]. Both folds perform differently.…”

Section: A Forest Areasmentioning

confidence: 99%

“…An additional miscelaneous class exists for partial objects and background scatter. We follow the experimental setup by [17] and remove that class from the evaluation. Furthermore, we adapt their splits for training and testing sets in our initial experiment.…”

Section: A Forest Areasmentioning

confidence: 99%

“…1) Default Setting: For evaluation of semantic segmentation quality, we follow [17] in using the common criteria precision, recall and F1-score. The latter is the harmonic mean of the first two performance measures.…”

Section: A Forest Areasmentioning

confidence: 99%

“…5. In addition to obtaining best performance, our method has a lower computational effort than the method described in [17]. It does not need to compute features for every single point.…”

Section: A Forest Areasmentioning

confidence: 99%

See 4 more Smart Citations

Region-Based Edge Convolutions With Geometric Attributes for the Semantic Segmentation of Large-Scale 3-D Point Clouds

Contreras¹,

Sickert

Denzler

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

In this paper, we present a semantic segmentation framework for large-scale 3D point clouds with high spatial resolution. For such data with huge amounts of points, the classification of each individual 3D point is an intractable task. Instead, we propose to segment the scene into meaningful regions as a first step. Afterward, we classify these segments using a combination of PointNet and geometric deep learning. This two-step approach resembles object-based image analysis. As an additional novelty, we apply surface normalization techniques and enrich features with geometric attributes. Our experiments show the potential of this approach for a variety of outdoor scene analysis tasks. In particular, we are able to reach 89.6% overall accuracy and 64.4% average intersection over union (IoU) in the Semantic3D benchmark. Furthermore, we achieve 66.7% average IoU on Paris-Lille-3D. We also successfully apply our approach to the automatic semantic analysis of forestry data.

show abstract

Section: B Points-wise Approachesmentioning

confidence: 99%

Section: A Forest Areasmentioning

confidence: 99%

Section: A Forest Areasmentioning

confidence: 99%

Section: A Forest Areasmentioning

confidence: 99%

“…5. In addition to obtaining best performance, our method has a lower computational effort than the method described in [17]. It does not need to compute features for every single point.…”

Section: A Forest Areasmentioning

confidence: 99%

See 3 more Smart Citations

Region-Based Edge Convolutions With Geometric Attributes for the Semantic Segmentation of Large-Scale 3-D Point Clouds

Contreras¹,

Sickert

Denzler

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

Edge-Convolution Point Net for Semantic Segmentation of Large-Scale Point Clouds

Contreras

Denzler

2019

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

View full text Add to dashboard Cite

Automatically Annotated Dataset of a Ground Mobile Robot in Natural Environments via Gazebo Simulations

Sánchez

Morales

Martínez

et al. 2022

Sensors

View full text Add to dashboard Cite

This paper presents a new synthetic dataset obtained from Gazebo simulations of an Unmanned Ground Vehicle (UGV) moving on different natural environments. To this end, a Husky mobile robot equipped with a tridimensional (3D) Light Detection and Ranging (LiDAR) sensor, a stereo camera, a Global Navigation Satellite System (GNSS) receiver, an Inertial Measurement Unit (IMU) and wheel tachometers has followed several paths using the Robot Operating System (ROS). Both points from LiDAR scans and pixels from camera images, have been automatically labeled into their corresponding object class. For this purpose, unique reflectivity values and flat colors have been assigned to each object present in the modeled environments. As a result, a public dataset, which also includes 3D pose ground-truth, is provided as ROS bag files and as human-readable data. Potential applications include supervised learning and benchmarking for UGV navigation on natural environments. Moreover, to allow researchers to easily modify the dataset or to directly use the simulations, the required code has also been released.

show abstract

Semantic Segmentation of Outdoor Areas Using 3D Moment Invariants and Contextual Cues

Cited by 3 publications

References 20 publications

Region-Based Edge Convolutions With Geometric Attributes for the Semantic Segmentation of Large-Scale 3-D Point Clouds

Region-Based Edge Convolutions With Geometric Attributes for the Semantic Segmentation of Large-Scale 3-D Point Clouds

Edge-Convolution Point Net for Semantic Segmentation of Large-Scale Point Clouds

Automatically Annotated Dataset of a Ground Mobile Robot in Natural Environments via Gazebo Simulations

Contact Info

Product

Resources

About