Structure SLAM with points, planes and objects

Zhou, Benchun; Gilles, Maximilian; Meng, Yongqi

doi:10.1080/01691864.2022.2123253

Cited by 6 publications

(9 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Contrasting with the plane extraction technique in [ 27 , 28 ], the method proposed herein capitalizes on known map information to extract plane point clouds. The method introduced in [ 30 ] should meet the criteria that planes be parallel or perpendicular to one another, a strict criterion that restricts localization application. Within numerous well-designed structures, corridors are not always aligned horizontally and vertically.…”

Section: Discussionmentioning

confidence: 99%

“…[ 29 ] utilized the absolute ground plane to constrain vertical pose estimation, subsequently reducing the estimation error. A point-plane-object localization system was proposed for semantic map reconstruction [ 30 ], which demonstrated effective localization in indoor scenarios. However, the method in [ 30 ] necessitates strict criteria, such as planes being parallel or perpendicular to one another.…”

Section: Introductionmentioning

confidence: 99%

“…A point-plane-object localization system was proposed for semantic map reconstruction [ 30 ], which demonstrated effective localization in indoor scenarios. However, the method in [ 30 ] necessitates strict criteria, such as planes being parallel or perpendicular to one another. Additionally, in both [ 27 , 28 ], planes are identified solely by analyzing point clouds.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Localization on a-priori information of plane extraction

Zhao

Misyurin

et al. 2023

PLoS ONE

View full text Add to dashboard Cite

Localization constitutes a critical challenge for autonomous mobile robots, with flattened walls serving as a fundamental reference for indoor localization. In numerous scenarios, prior knowledge of a wall’s surface plane is available, such as planes in building information modeling (BIM) systems. This article presents a localization technique based on a-priori plane point cloud extraction. The position and pose of the mobile robot are estimated through real-time multi-plane constraints. An extended image coordinate system is proposed to represent any planes in space and establish correspondences between visible planes and those in the world coordinate system. Potentially visible points representing the constrained plane in the real-time point cloud are filtered using the filter region of interest (ROI), derived from the theoretical visible plane region within the extended image coordinate system. The number of points representing the plane influences the calculation weight in the multi-plane localization approach. Experimental validation of the proposed localization method demonstrates its allowance for redundancy in initial position and pose error.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Localization on a-priori information of plane extraction

Zhao

Misyurin

et al. 2023

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…MVPI can effectively fuse the 3D corner information based on 2D feature regression and the depth information based on depth feature map, so as to improve the accuracy of the final depth estimation. Finally, we tested MonoSPDC on KITTI [15] dataset. The test results show that MonoSPDC can achieve the accuracy of Transformer structure while ensuring the speed advantage of convolution operation.…”

Section: Introductionmentioning

confidence: 99%

MonoSPDC：A Novel Monocular 3D Object Detection Using Global-Channel Atten- tion Optimization

Su,

Guo,

Liu

et al. 2024

Preprint

View full text Add to dashboard Cite

With the development of autonomous driving and mobile robotics, monocular 3D object detection has attracted attention in the above fields due to its advantages of low cost, fast speed and low power consumption. However, both mainstream monocular 3D object detection network structures currently face difficulties in practical application in these fields. One issue is the low accuracy of convolutional network structures. Another issue is the complex structure, high computational complexity, large training dataset, and LiDAR assisted training problem of the Transformer network structure. To solve the above problems, a novel monocu-lar 3D object detection algorithm named MonoSPDC is proposed in this paper. In MonoSPDC, we propose the Feature Global-Channels Recombination Module (FGCR). FGCR uses the attention mechanism based on convolutional structure to perform attention operations on the global and channel dimensions of the feature map. This can make the feature map more distinct, while also making the generation of heatmap more reliable and accurate. we also propose the Multi-View Predictive Integration module (MVPI) in MonoSPDC. MVPI can realize the uncertain fusion of 3D corner information based on 2D feature regression and depth information based on depth feature map, so as to improve the accuracy of the final depth estimation. In summary, proposed Mono-SPDC does not require other additional auxiliary training data. It can retain the speed advantage of the convo-lutional structure while achieving the accuracy of the Transformer structure. In the test on the public dataset KITTI, our proposed MonoSPDC has achieved very excellent results.

show abstract

“…Convolutional Neural Networks (CNNs) have always been a concern in artificial intelligence. Since the AlexNet [1], ResNet [2] convolutional networks have been widely proven to be effective in object detection [3–9], semantic segmentation [10–15], and image classification etc. Recently, attention mechanisms in deep networks have received extensive attention, which stems from the studies on human vision and provides an efficient way for us to work.…”

Section: Introductionmentioning

confidence: 99%

An efficient mixed attention module

Sheng

Chen

2023

IET Computer Vision

View full text Add to dashboard Cite

Recently, the application of attention mechanisms in convolutional neural networks (CNNs) has become a hot area in computer vision. Most existing methods focus on channel attention or spatial attention. Some mixed attention usually achieves better performance than channel attention or spatial attention with the help of a complex model structure, which increases the complexity of the model. This article proposes an efficient mixed attention that combines channel information with spatial information using learnable broadcast addition to reduce this complexity. In particular, this module can simplify learning and improve performance with fewer parameters. Furthermore, our method uses an excitation method based on the Tanh function to reduce computational resources while maintaining model performance, and it is a lightweight attention module that can be used in arbitrary CNNs to improve performance. Experiments on ImageNet and Cifar confirm the effectiveness of the proposed method. Besides, our method remains highly competitive for object detection tasks and image segmentation tasks.

show abstract

Structure SLAM with points, planes and objects

Cited by 6 publications

References 32 publications

Localization on a-priori information of plane extraction

Localization on a-priori information of plane extraction

MonoSPDC：A Novel Monocular 3D Object Detection Using Global-Channel Atten- tion Optimization

An efficient mixed attention module

Contact Info

Product

Resources

About