Superquadric Object Representation for Optimization-based Semantic SLAM

Tschopp, Florian; Nieto, Juan; Siegwart, Roland; Cadena, Cesar

doi:10.48550/arxiv.2109.09627

Cited by 2 publications

(5 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, our training data can be optimized by simulating more realistic semantic object distributions or by fine-tuning on annotated real-world data. Our method could also be combined with other semantic object representations for better object matching results [11].…”

Section: Discussionmentioning

confidence: 99%

“…Dynamic and unknown objects are removed in pre-processing. We represent the objects using a 3D point located in the object's centroid [11] and reformulate the data format as (x, y, z, instance id, class). Tab.…”

Section: Global Localization Performancementioning

confidence: 99%

“…Object-based descriptors are proposed to solve the problem by using high-level representations, such as lines [9], cubes [10], and super quadrics [11]. Semantic information can also be attached to uniquely distinguish each object.…”

Section: Introductionmentioning

confidence: 99%

“…Semantic information can also be attached to uniquely distinguish each object. Unlike appearance-based descriptors, instances can be robustly detected in the presence of changing perspectives, illumination, and object occlusion [11]. Some proposed For each object in the global and local map, we generate a graph by specifying an edge between every pair of neighboring nodes according to the Euclidean distance.The graph is used to extract robust descriptors using GCNs.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Descriptellation: Deep Learned Constellation Descriptors for SLAM

Xing¹,

Sun²,

Cramariuc³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Current global localization descriptors in Simultaneous Localization and Mapping (SLAM) often fail under vast viewpoint or appearance changes. Adding topological information of semantic objects into the descriptors ameliorates the problem. However, handcrafted topological descriptors extract limited information and they are not robust to environmental noise, drastic perspective changes, or object occlusion or misdetections. To solve this problem, we formulate a learningbased approach by constructing constellations from semantically meaningful objects and use Deep Graph Convolution Networks to map the constellation representation to a descriptor. We demonstrate the effectiveness of our Deep Learned Constellation Descriptor (Descriptellation) on the Paris-Rue-Lille and IQmulus datasets. Although Descriptellation is trained on randomly generated simulation datasets, it shows good generalization abilities on real-world datasets. Descriptellation outperforms the PointNet and handcrafted constellation descriptors for global localization, and shows robustness against different types of noise.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Global Localization Performancementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Descriptellation: Deep Learned Constellation Descriptors for SLAM

Xing¹,

Sun²,

Cramariuc³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Location accuracy is one of the most basic assessment standards in the SLAM system and is a precondition for mobile robots to perform many tasks [225]. Introducing environmental semantic information can effectively improve the scale uncertainty and cumulative drift in visual SLAM localization, thus improving the localization accuracy to varying degrees [226].…”

Section: Semantic With Locationmentioning

confidence: 99%

An Overview on Visual SLAM: From Tradition to Semantic

et al. 2022

View full text Add to dashboard Cite

Visual SLAM (VSLAM) has been developing rapidly due to its advantages of low-cost sensors, the easy fusion of other sensors, and richer environmental information. Traditional visionbased SLAM research has made many achievements, but it may fail to achieve wished results in challenging environments. Deep learning has promoted the development of computer vision, and the combination of deep learning and SLAM has attracted more and more attention. Semantic information, as high-level environmental information, can enable robots to better understand the surrounding environment. This paper introduces the development of VSLAM technology from two aspects: traditional VSLAM and semantic VSLAM combined with deep learning. For traditional VSLAM, we summarize the advantages and disadvantages of indirect and direct methods in detail and give some classical VSLAM open-source algorithms. In addition, we focus on the development of semantic VSLAM based on deep learning. Starting with typical neural networks CNN and RNN, we summarize the improvement of neural networks for the VSLAM system in detail. Later, we focus on the help of target detection and semantic segmentation for VSLAM semantic information introduction. We believe that the development of the future intelligent era cannot be without the help of semantic technology. Introducing deep learning into the VSLAM system to provide semantic information can help robots better perceive the surrounding environment and provide people with higher-level help.

show abstract

Superquadric Object Representation for Optimization-based Semantic SLAM

Cited by 2 publications

References 38 publications

Descriptellation: Deep Learned Constellation Descriptors for SLAM

Descriptellation: Deep Learned Constellation Descriptors for SLAM

An Overview on Visual SLAM: From Tradition to Semantic

Contact Info

Product

Resources

About