检索结果-内蒙古大学图书馆

Adaptive multimodal fusion with semantic priority scheduling for cooperative 3d detection and communication methods

ALEXANdRIA ENGINEERING JOURNAL 2025年 124卷 362-375页

作者： Wang, Feng dong, Meixia Zou, Jiajun Ye, Shitong Wan, Zhiping Liu, Shaojiang Guangzhou Xinhua Univ Sch Informat & Intelligent Engn Dongguan 523133 Peoples R China Guangzhou Huashang Coll Artificial Intelligence Inst Guangzhou 511300 Peoples R China

In response to the urgent demand for high-precision 3d object detection and efficient communication in intelligent transportation systems and autonomous driving, this paper proposes an Adaptive Multimodal Fusion and Semantic Priority Scheduling approach for collaborative 3d detection and communication (AMFS-C3d). during the fusion of multimodal sensor data (from cameras, LidAR, and radar), a graph neural network is introduced to structurally encode the integrated 3d features, thoroughly capturing spatial interactions among targets;a variational autoencoder then compresses these high-dimensional graph embeddings, retaining essential semantic information even under bandwidth constraints and significant channel noise. To further reduce latency and increase communication efficiency, the system employs an adaptive priority scheduling mechanism that dynamically allocates bandwidth based on real-time network load and target importance, ensuring the timely transmission of critical objects. Experimental results show that AMFS-C3d significantly outperforms comparative methods in key metrics such as mean Average Precision (mAP), recall, semantic fidelity, and transmission latency. Under the same mAP threshold, AMFS-C3d reduces bandwidth requirements by over 10 % on average, and under high signal-to-noise ratio conditions, it improves detection accuracy and recall by 3-5 percentage points. Moreover, AMFS-C3d demonstrates superior applicability and robustness across diverse network loads and channel conditions.

关键词： Semantic communication 3d detection Multimodal data fusion Graph neural networks Adaptive priority scheduling

来源：评论

学校读者我要写书评

暂无评论

3d detection of internal defects for gas turbine blades

引用

INSIGHT 2017年第7期59卷 364-370页

作者： Li, Bing Chen, Lei Wang, Yuegen Gao, Mengqiu Xi An Jiao Tong Univ State Key Lab Mfg Syst Engn Xian 710049 Shaanxi Peoples R China Xi An Jiao Tong Univ Collaborat Innovat Ctr High End Mfg Equipment Xian 710049 Shaanxi Peoples R China Xi An Jiao Tong Univ Int Joint Res Lab Micro Nano Mfg & Measurement Te Xian 710049 Shaanxi Peoples R China

The blades are an important part of gas turbines and internal defects, such as cracks and holes, can occur in some gas turbine blades. due to the harsh working environment, these defects have a great influence on the safe operation of the gas turbine in practical applications. In this paper, a new method based on a digital radiography (dR) system is adopted to detect defects in gas turbine blades. Considering the limited size of the flat-panel detector in a dR system and the uneven thickness distribution of the irregular-shaped gas turbine blades, a region segmentation-based inspection method is presented in this paper. That is, the gas turbine blade is divided into a few regions using lead wires and the dR system is then used to inspect these sub-regions. Considering the high cost and low efficiency limitations of high-power computed tomography (CT) and traditional film-based radiographic testing (RT), certain image processing methods are adopted to accurately automate the extraction and analysis processes. A cubic polynomial interpolation background subtraction method is adopted for the accurate extraction of defects. In addition, a finite element method is presented to obtain the 3d parameter (the volume) of a defect from its 2d radiographic image. Experimental results show that the proposed method can efficiently estimate the 2d and 3d geometric parameters of defects in gas turbine blades at a low cost. The method is particularly suitable for offline X-ray inspection of gas turbine blades.

关键词： gas turbine blades defects 3d detection region segmentation image processing

来源：评论

学校读者我要写书评

暂无评论

Inter-Vehicle distance Estimation Method Based on Monocular Vision Using 3d detection

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2020年第5期69卷 4907-4919页

作者： Zhe, Ting Huang, Liqin Wu, Qiang Zhang, Jianjia Pei, Chenhao Li, Liangyu Fuzhou Univ Coll Phys & Informat Engn Fuzhou 350108 Peoples R China Univ Technol Sydney Sch Elect & Data Engn Sydney NSW 2007 Australia

Most autonomous vehicles build their perception systems on expensive sensors, such as LIdAR, RAdAR, and high-precision Global Positioning System (GPS). However, cameras can provide richer sensing at a considerably lower cost, this makes them a more appealing alternative. A driving assistance system (dAS) based on monocular vision has gradually become a research hotspot, and inter-vehicle distance estimation based on monocular vision is an important technology in dAS. There are still constrains in the existing methods for estimating the inter-vehicle distance based on monocular vision, such as low accuracy when distance is larger, unstable accuracy for different types vehicles, and significantly poor performance on distance estimation for severely occluded vehicles. To improve the accuracy and robustness of ranging results, this study proposes a monocular vision end-to-end inter-vehicle distance estimation method based on 3d detection. The actual area of the rare view of the vehicle and the corresponding projection area in the image are obtained by 3d detection method. An area-distance geometric model is then established on the basis of the camera projection principle to recover distance. Our method shows its potential in complex traffic scenarios by testing the test set data provided on the real-world computer vision benchmark, KITTI. The experimental results have superior performance than the existing published methods. Moreover, the accuracy of occluded vehicle ranging results can reach approximately 98%, while the accuracy deviation between vehicles with different visual angles is less than 2%.

关键词： Monocular vision inter-vehicle distance estimation area-distance geometric model 3d detection

来源：评论

学校读者我要写书评

暂无评论

Accurate and Efficient LIF-Nets for 3d detection and Recognition

引用

IEEE ACCESS 2020年 8卷 98562-98571页

作者： Shi, Yueting Li, Hai Zhang, Hehui Wu, Zhenzhi Ren, Shiwei Beijing Inst Technol Sch Informat & Elect Beijing 100081 Peoples R China Beijing Univ Posts & Telecommun Sch Software Engn Beijing 100876 Peoples R China Beijing Lynxi Technol Co Ltd Beijing 100080 Peoples R China

3d object detection and recognition are crucial tasks for many spatiotemporal processing applications, such as computer-aided diagnosis and autonomous driving. Although prevalent 3d Convolution Nets (ConvNets) have continued to improve the accuracy and sensitivity, excessive computing resources are required. In this paper, we propose Leaky Integrate and Fire Networks (LIF-Nets) for 3d detection and recognition tasks. LIF-Nets have rich inter-frame sensing capability brought from membrane potentials, and low power event-driven mechanism, which make them excel in 3d processing and save computational cost at the same time. We also develop ResLIF Blocks to solve the degradation problem of deep LIF-Nets, and employ U-LIF structure to improve the feature representation capability. As a result, we carry out experiments on the LUng Nodule Analysis 2016 (LUNA16) public dataset for chest CT automated analysis and conclude that the LIF-Nets achieve 94.6% detection sensitivity at 8 False Positives per scan and 94.14% classification accuracy while the LIF-detection net reduces 65.45% multiplication operations, 65.12% addition operations, and 65.32% network parameters. The results show that LIF-Nets have extraordinary time-efficient and energy-saving performance while achieving comparable accuracy.

关键词： Three-dimensional displays Computed tomography Feature extraction Object detection Spatiotemporal phenomena Spiking neural network leaky integrate and fire model 3d detection 3d recognition pulmonary nodule screening

来源：评论

学校读者我要写书评

暂无评论

ACF-Net: Asymmetric Cascade Fusion for 3d detection With LidAR Point Clouds and Images

IEEE TRANSACTIONS ON INTELLIGENT VEHICLES

引用

IEEE TRANSACTIONS ON INTELLIGENT VEHICLES 2024年第2期9卷 3360-3371页

作者： Tian, Yonglin Zhang, Xianjing Wang, Xiao Xu, Jintao Wang, Jiangong Ai, Rui Gu, Weihao ding, Weiping Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Haomo Technol Co Ltd AI Ctr Beijing 100192 Peoples R China Anhui Univ Sch Artificial Intelligence Hefei 230031 Peoples R China Haomo Technol Co Ltd Beijing 100192 Peoples R China Nantong Univ Sch Informat Sci & Technol Nantong 226019 Peoples R China

The recognition and utilization of complementary information arising from modality-intrinsic properties play crucial roles in multimodal 3d detection. However, most of the current approaches for fusion-based 3d detection follow symmetrical fusion paradigms and adopt early fusion, middle fusion as well as late fusion styles, which ignore the unequal status of data with different modalities. In this paper, according to the timing of fusion, we adopt an asymmetric cascade fusion network to exploit both the structural information from point clouds and the complementary semantic information from images. A multi-stage cascade design of 3d object detection is proposed to iteratively refine predictions and several late image features (comprised of detection clues, segmentation clues, and deep features from encoders) are incorporated into different stages of the LidAR branch to maintain the integrity of image features and enable deep multimodal interactions. Besides, to mitigate the effects of the down-sampling of voxelized features and possible mismatching of multimodal data, we propose proxy-based cross-modality sampling to utilize the high-density point clouds coordinates and develop an image degeneration process to simulate the noise in cross-modality matching for robust training. Extensive experiments are conducted on KITTI and Waymo Open dataset, which validate the effectiveness of the proposed method.

关键词： Three-dimensional displays Feature extraction Point cloud compression Laser radar Object detection Timing Fuses 3d detection autonomous driving asymmetric fusion cascade fusion multimodal fusion

来源：评论

学校读者我要写书评

暂无评论

A low-cost pipeline surface 3d detection method used on robots

引用

SIGNAL IMAGE ANd VIdEO PROCESSING 2024年第4期18卷 3459-3470页

作者： Lan, Tianxiang Yang, Guotian North China Elect Power Univ Sch Control & Engn Beijing 102206 Peoples R China

This paper presents a low-cost pipeline surface 3d detection method based on line structured light. The method adds 3d detection to the robot at a meager cost and change. The optical flow method is used to derive the motion information of the robot to replace the motion closed loop. Finally, the depth data for the entire surface are generated automatically only from the camera and the line laser projector, without using other devices. In addition, a novel spot centroid extraction algorithm based on the color region of interest and an adaptive threshold is presented. This method can accurately detect the laser centroid in the pipeline surface. We conducted experiments with a quadruped robot and validated algorithms. The experimental results show that the proposed method achieves 3d detection on a trackless robot at a meager cost and is superior to standard depth sensors.

关键词： 3d detection Line structured light Optical flow Laser Robot

来源：评论

学校读者我要写书评

暂无评论

GVnet: Gaussian model with voxel-based 3d detection network for autonomous driving

引用

NEURAL COMPUTING & APPLICATIONS 2022年第9期34卷 6637-6645页

作者： Qin, Peilin Zhang, Chuanwei dang, Meng Xian Univ Sci & Technol Sch Mech Engn Xian 710054 Shaanxi Peoples R China

This paper proposed a two-stage Voxel-based 3d Object detector which named GVnet. Voxel-based method mainly relies on sampling and Grouping point in voxel and the feature map generated by subsequent 3d CNN to control the quality of detection. Moreover, traditional voxel feature encoder (VFE) methods cannot adjust the quality of feature map through reasonable sampling. Therefore, the method we propose is an improvement to the existing VFE. The specific operations are: First calculate the corresponding Gaussian distribution of the original point cloud data, and then sampling any number of points by controlling the confidence value to improve the performance of voxel encoder and further improve the quality of the feature map output by the 3d CNN. In addition, a voxel ROI pooling method is proposed in stage two. In ROI Pooling, the receptive field in the original space and the corresponding raw point are obtained through the mapping relationship between feature and ROI, then change the raw point to adjust the receptive field to improve the performance of classification and regression. Finally, the experimental results on the KITTI, nuScenes and Waymo dataset show that the performance of GVnet under most of the evaluation indexes is better than the current detection methods, at the cost of only a small amount of inference time.

关键词： 3d detection Voxel-based detector Gaussian-voxel feature encoding Voxel-ROI pooling

来源：评论

学校读者我要写书评

暂无评论

3d detection and Reconstruction of Worn Parts for Flexible Remanufacture

3D Detection and Reconstruction of Worn Parts for Flexible R...

引用

3rd International Conference on Manufacturing Science and Engineering (ICMSE 2012)

作者： Yin, Ziqiang Zhang, Guangjun Yuan, Xin Zhao, Huihui Wu, Lin 28 Zhejiang Rd Qingdao Peoples R China

ISBN: (纸本)9783037853696

In recent years, flexible remanufacture has become new research focus because of the requirement of rapid repair and reuse worn parts. 3d modeling of the worn parts is the precondition of flexible remanufacturing system. According to the random position and the various forms of worn parts, 3d detection and reconstruction are necessary. First, this paper establishes a 3d detection and reconstruction system based on linear laser. The mathematic model, work principle and composition of this system are introduced, and then, according to the analysis of this system, a two steps rapid calibration algorithm is proposed. When the system parameters are calibrated, it is able to acquire 3d information on surface of worn parts in the form of scattered point data clouds. Finally, take a coupler kunckles as the worn part, 3d detection and reconstruction verifying experiment is carried out.

关键词： 3d detection reconstruction remanufacture structured light calibration robotic arc welding

来源：评论

学校读者我要写书评

暂无评论

Automated 3d detection and classification of Giardia lamblia cysts using digital holographic microscopy with partially coherent source

Automated 3D detection and classification of Giardia lamblia...

引用

Conference on Optical Modelling and design II

作者： El Mallahi, A. detavernier, A. Yourassowsky, C. dubois, F. Univ Libre Bruxelles Micrograv Res Ctr CP 165-6250 Ave Franklin Roosevelt B-1050 Brussels Belgium Univ Libre Bruxelles Dept Water Pollut Control B-1050 Brussels Belgium

ISBN: (纸本)9780819491213

Over the past century, monitoring of Giardia lamblia became a matter of concern for all drinking water suppliers worldwide. Indeed, this parasitic flagellated protozoan is responsible for giardiasis, a widespread diarrhoeal disease (200 million symptomatic individuals) that can lead immunocompromised individuals to death. The major difficulty raised by Giardia lamblia's cyst, its vegetative transmission form, is its ability to survive for long periods in harsh environments, including the chlorine concentrations and treatment duration used traditionally in water disinfection. Currently, there is a need for a reliable, inexpensive, and easy-to-use sensor for the identification and quantification of cysts in the incoming water. For this purpose, we investigated the use of a digital holographic microscope working with partially coherent spatial illumination that reduces the coherent noise. digital holography allows one to numerically investigate a volume by refocusing the different plane of depth of a hologram. In this paper, we perform an automated 3d analysis that computes the complex amplitude of each hologram, detects all the particles present in the whole volume given by one hologram and refocuses them if there are out of focus using a refocusing criterion based on the integrated complex amplitude modulus and we obtain the (x,y,z) coordinates of each particle. Then the segmentation of the particles is processed and a set of morphological and textures features characteristic to Giardia lamblia cysts is computed in order to classify each particles in the right classes.

关键词： digital holography Image processing 3d detection Classification

来源：评论

学校读者我要写书评

暂无评论

Large field of view 3d detection with a bionic curved compound-eye camera

Large field of view 3D detection with a bionic curved compou...

引用

Conference on AOPC - Optical Sensing and Imaging Technology

作者： Liu, Jinheng Zhang, Yuanjie Xu, Huangrong Yu, Weixing Chinese Acad Sci Xian Inst Opt & Precis Mech Key Lab Spectral Imaging Technol 17 Xinxi Rd Xian 710119 Peoples R China Univ Chinese Acad Sci Ctr Mech Sci & Optoelect Engn Beijing 100049 Peoples R China

ISBN: (纸本)9781510650060;9781510650053

In the bionic curved compound-eye camera (BCCEC) we have invented, the overlapping field of view (FOV) among the ommatidia makes 3d detection possible. In this work, we analyzed the overlapping FOV in BCCEC in detail to prove its potential in 3d detection and designed a new experiment to test its performance. In that the FOVs of multiple ommatidia in BCCEC overlap each other, the FOV of a single ommatidium is used as a representative analysis. The relationship between the overlapping ratio of FOV and the object distance is quantitatively calculated. The results show that more than 95% of the FOV can be 3d reconstructed when the object distance exceeds 32 cm. Next, in order to realize the automatic calibration of all ommatidia, ommatidia are numbered and an addressing algorithm based on the number information of ommatidia is designed, which can be used to determine the adjacent ommatidia of any ommatidium so as to acquire the ommatidia pairs that need to be calibrated. Then, since the aperture of ommatidium is relatively small and it is difficult to accurately align, a new 3d detection experiment is designed. The laser rangefinder is fixed, the black paper is used to block the laser and the formed light spot is used as the detection target. The experimental results show that 3d detection can be performed in the whole FOV of BCCEC. The BCCEC can obtain multi-dimensional information in a large FOV, and it have greater application potential in obstacle avoidance and navigation.

关键词： compound-eye FOV 3d detection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：