Search

Facing the significant challenge of 3D object detection in complex weather conditions and road environments, existing algorithms based on single-frame point cloud data struggle to achieve desirable results. These methods typically focus on spatial relationships within a single frame, overlooking the semantic correlations and spatiotemporal continuity between consecutive frames. This leads to discontinuities and abrupt changes in the detection outcomes. To address this issue, this paper proposes a multi-frame 3D object detection algorithm based on a deformable spatiotemporal Transformer. Specifically, a deformable cross-scale Transformer module is devised, incorporating a multi-scale offset mechanism that non-uniformly samples features at different scales, enhancing the spatial information aggregation capability of the output features. Simultaneously, to address the issue of feature misalignment during multi-frame feature fusion, a deformable cross-frame Transformer module is proposed. This module incorporates independently learnable offset parameters for different frame features, enabling the model to adaptively correlate dynamic features across multiple frames and improve the temporal information utilization of the model. A proposal-aware sampling algorithm is introduced to significantly increase the foreground point recall, further optimizing the efficiency of feature extraction. The obtained multi-scale and multi-frame voxel features are subjected to an adaptive fusion weight extraction module, referred to as the proposed mixed voxel set extraction module. This module allows the model to adaptively obtain mixed features containing both spatial and temporal information. The effectiveness of the proposed algorithm is validated on the KITTI, nuScenes, and self-collected urban datasets. The proposed algorithm achieves an average precision improvement of 2.1% over the latest multi-frame-based algorithms. [ABSTRACT FROM AUTHOR]

In recent years, 3D object detection based on LiDAR point clouds is a key component of autonomous driving. In pursuit of enhancing the accuracy of 3D point cloud feature extraction and point cloud detection, this paper introduces a novel 3D object detection model, termed as Graph Self-Attention-RCNN (GA-RCNN). This model is designed to integrate voxel information and point location information, enhancing the quality of 3D object proposals while maintaining contextual accuracy. The first stage rectifies the previous approach that relied on local features for preselected boxes, overlooking crucial global contextual information. An improved method is suggested in this work, utilizing BEV to capture long-range dependencies via a cross-attention mechanism. The second stage addresses the overreliance on local neighborhood point feature extraction. The Graph Self-Attention Pooling method is proposed, characterized by its dynamic computation of contribution weights for inputs. This enhances the model's flexibility and generalization performance. Extensive evaluations on KITTI and Waymo datasets demonstrate GA-RCNN's superior accuracy compared to other methods, affirming its efficacy in 3D object detection. [ABSTRACT FROM AUTHOR]

Three-dimensional object detection plays a crucial role in achieving accurate and reliable autonomous driving systems. However, the current state-of-the-art two-stage detectors lack flexibility and have limited feature extraction capabilities to effectively handle the disorder and irregularity of point clouds. In this paper, we propose a novel network called FANet, which combines the strengths of PV-RCNN and PAConv (position adaptive convolution). The goal of FANet is to address the irregularity and disorder present in point clouds. In our network, the convolution operation constructs convolutional kernels using a basic weight matrix, and the coefficients of these kernels are adaptively learned by LearnNet from relative points. This approach allows for the flexible modeling of complex spatial variations and geometric structures in 3D point clouds, leading to the improved extraction of point cloud features and generation of high-quality 3D proposal boxes. Compared to other methods, extensive experiments on the KITTI dataset have demonstrated that the FANet exhibits superior 3D object detection accuracy, showcasing a significant improvement in our approach. [ABSTRACT FROM AUTHOR]

Three-dimensional object detection in the point cloud can provide more accurate object data for autonomous driving. In this paper, we propose a method named MA-MFFC that uses an attention mechanism and a multi-scale feature fusion network with ConvNeXt module to improve the accuracy of object detection. The multi-attention (MA) module contains point-channel attention and voxel attention, which are used in voxelization and 3D backbone. By considering the point-wise and channel-wise, the attention mechanism enhances the information of key points in voxels, suppresses background point clouds in voxelization, and improves the robustness of the network. The voxel attention module is used in the 3D backbone to obtain more robust and discriminative voxel features. The MFFC module contains the multi-scale feature fusion network and the ConvNeXt module; the multi-scale feature fusion network can extract rich feature information and improve the detection accuracy, and the convolutional layer is replaced with the ConvNeXt module to enhance the feature extraction capability of the network. The experimental results show that the average accuracy is 64.60% for pedestrians and 80.92% for cyclists on the KITTI dataset, which is 1.33% and 2.1% higher, respectively, compared with the baseline network, enabling more accurate detection and localization of more difficult objects. [ABSTRACT FROM AUTHOR]

Cooperative perception in the field of connected autonomous vehicles (CAVs) aims to overcome the inherent limitations of single-vehicle perception systems, including long-range occlusion, low resolution, and susceptibility to weather interference. In this regard, we propose a high-precision 3D object detection V2V cooperative perception algorithm. The algorithm utilizes a voxel grid-based statistical filter to effectively denoise point cloud data to obtain clean and reliable data. In addition, we design a feature extraction network based on the fusion of voxels and PointPillars and encode it to generate BEV features, which solves the spatial feature interaction problem lacking in the PointPillars approach and enhances the semantic information of the extracted features. A maximum pooling technique is used to reduce the dimensionality and generate pseudoimages, thereby skipping complex 3D convolutional computation. To facilitate effective feature fusion, we design a feature level-based crossvehicle feature fusion module. Experimental validation is conducted using the OPV2V dataset to assess vehicle coperception performance and compare it with existing mainstream coperception algorithms. Ablation experiments are also carried out to confirm the contributions of this approach. Experimental results show that our architecture achieves lightweighting with a higher average precision (AP) than other existing models. [ABSTRACT FROM AUTHOR]

Object detection plays an important role in autonomous driving systems. LiDAR is widely used in autonomous driving vehicles and robots as a sensor for environmental perception. Recently, the development of computational power and deep learning technology makes it possible to classify and locate objects from LiDAR point cloud in a single end-to-end learnable network. However, objects are sparsely distributed in large point cloud field, and are always been partly scanned by LiDAR, which pose a challenge for accurate and rapid object positioning and classification from the raw point cloud. In this paper, we introduce a new single-shot refinement neural network for fast and accurate 3D object detection from the raw LiDAR point cloud. Firstly, we exploit self-attention mechanism in main object detection branch to enhance object feature representation. Secondly, we apply deformable convolution for learning adaptive receptive fields to fully capture the features of rotating and partially visible objects. Thirdly, an object refinement branch is introduced to produce a finer regression of objects upon the primary estimation from the main detection branch. All proposed modules have been proven to effectively improve the accuracy of object detection. Our method is evaluated on KITTI 3D detection benchmark and achieves state-of-the-art results while maintains real-time efficiency. Furthermore, real-time test in autonomous driving vehicle demonstrates that our method is robust to 16 channels LiDAR and can meet the demands of accuracy, efficiency, and visibility of object detection in various urban scenarios. [ABSTRACT FROM AUTHOR]

Current state-of-the-art (SOTA) LiDAR-only detectors perform well for 3D object detection tasks, but point cloud data are typically sparse and lacks semantic information. Detailed semantic information obtained from camera images can be added with existing LiDAR-based detectors to create a robust 3D detection pipeline. With two different data types, a major challenge in developing multi-modal sensor fusion networks is to achieve effective data fusion while managing computational resources. With separate 2D and 3D feature extraction backbones, feature fusion can become more challenging as these modes generate different gradients, leading to gradient conflicts and suboptimal convergence during network optimization. To this end, we propose a 3D object detection method, Attention-Enabled Point Fusion (AEPF). AEPF uses images and voxelized point cloud data as inputs and estimates the 3D bounding boxes of object locations as outputs. An attention mechanism is introduced to an existing feature fusion strategy to improve 3D detection accuracy and two variants are proposed. These two variants, AEPF-Small and AEPF-Large, address different needs. AEPF-Small, with a lightweight attention module and fewer parameters, offers fast inference. AEPF-Large, with a more complex attention module and increased parameters, provides higher accuracy than baseline models. Experimental results on the KITTI validation set show that AEPF-Small maintains SOTA 3D detection accuracy while inferencing at higher speeds. AEPF-Large achieves mean average precision scores of 91.13, 79.06, and 76.15 for the car class's easy, medium, and hard targets, respectively, in the KITTI validation set. Results from ablation experiments are also presented to support the choice of model architecture. [ABSTRACT FROM AUTHOR]

In this paper, we propose a 3D object detection method called MMAF-Net that is based on the multi-view and multi-stage adaptive fusion of RGB images and LiDAR point cloud data. This is an end-to-end architecture, which combines the characteristics of RGB images, the front view of point clouds based on reflection intensity, and the bird's eye view of point clouds. It also adopts a multi-stage fusion approach of "data-level fusion + feature-level fusion" to fully exploit the strength of multimodal information. Our proposed method addresses key challenges found in current 3D object detection methods for autonomous driving, including insufficient feature extraction from multimodal data, rudimentary fusion techniques, and sensitivity to distance and occlusion. To ensure the comprehensive integration of multimodal information, we present a series of targeted fusion methods. Firstly, we propose a novel input form that encodes dense point cloud reflectivity information into the image to enhance its representational power. Secondly, we design the Region Attention Adaptive Fusion module utilizing an attention mechanism to guide the network in adaptively adjusting the importance of different features. Finally, we extend the 2D DIOU (Distance Intersection over Union) loss function to 3D and develop a joint regression loss based on 3D_DIOU and SmoothL1 to optimize the similarity between detected and ground truth boxes. The experimental results on the KITTI dataset demonstrate that MMAF-Net effectively addresses the challenges posed by highly obscured or crowded scenes while maintaining real-time performance and improving the detection accuracy of smaller and more difficult objects that are occluded at far distances. • A novel object detection method is proposed integrating RGB images and LiDAR data. • This study tests the method on the KITTI dataset. • This method tackles highly obscured or crowded scenes for autonomous driving. • This method improves the detection accuracy of smaller and occluded targets far away. [ABSTRACT FROM AUTHOR]

Accurate and comprehensive 3D objects detection is important for perception systems in autonomous driving. Nevertheless, contemporary mainstream methods tend to perform more effectively on large objects in regions proximate to the LiDAR, leaving limited exploration of long-range objects and small objects. The divergent point pattern of LiDAR, which results in a reduction in point density as the distance increases, leads to a non-uniform point distribution that is ill-suited to discretized volumetric feature extraction. To address this challenge, we propose the Foreground Voxel Proposal (FVP) module, which effectively locates and generates voxels at the foreground of objects. The outputs are subsequently merged to mitigating the difference in point cloud density and completing the object shape. Furthermore, the susceptibility of small objects to occlusion results in the loss of feature space. To overcome this, we propose the Multi-Scale Feature Integration Network (MsFIN), which captures contextual information at different ranges. Subsequently, the outputs of these features are integrated through a cascade framework based on transformers in order to supplement the object features space. The extensive experimental results demonstrate that our network achieves remarkable results. Remarkably, our approach demonstrated an improvement of 8.56% AP on the SECOND baseline for the Car detection task at a distance of more than 20 m, and 9.38% AP on the Cyclist detection task. [ABSTRACT FROM AUTHOR]

Three-dimensional object detection is the core of an autonomous driving perception system, which detects and analyzes targets around the vehicle to obtain their sizes, shapes, and categories to provide reliable operational decisions for achieving autonomous driving. To improve the detection and localization accuracy of multi-object targets such as surrounding vehicles and pedestrians in autonomous driving scenarios, based on PointPillars fast object detection network, a three-dimensional object detection algorithm based on the channel attention mechanism, ECA Modules-PointPillars, is proposed. Firstly, the improved algorithm uses point cloud columnarization features to convert a three-dimensional point cloud into a two-dimensional pseudo-image. Then, combining the 2D backbone network for feature extraction with the Efficient Channel Attention (ECA) modules to achieve the enhancement of the positional feature information in the pseudo-image and the weakening of the irrelevant feature information such as background noise. Finally, the single-shot multibox detector (SSD) algorithm was used to complete the 3D object detection task. The experimental results show that the improved algorithm improves the mAP by 3.84% and 4.04% in BEV mode and 3D mode, respectively, compared to PointPillars, which improves the mAP by 4.64% and 5.89% in BEV mode and 3D mode, respectively, compared to F-PointNet, improves the mAP by 11.78% and 14.19% in BEV mode and 3D mode, respectively, compared to VoxelNet, and improves the mAP by 9.47% and 6.55% in BEV mode and 3D mode, respectively, compared to SECOND, demonstrating the effectiveness and reliability of the improved algorithms in autonomous driving scenarios. [ABSTRACT FROM AUTHOR]

Accurately detect vehicles or pedestrians from 3D point clouds (3D object detection) is a fast developing research topic in autonomous driving and other domains. The fundamental component for feature extraction in 3D object detection is Set Abstraction (SA), which can downsample points while aggregating points to extract features. However, the current SA ignores the geometric and semantic properties of point clouds and may miss to detect remote small objects. In this paper, FocusSA is proposed, which consists two modules for enhancing useful feature extraction in the SA layer to improve 3D object detection accuracy. At first, Focused FPS (FocFPS) is proposed to evaluate the foreground and boundary scores of the points and reweighs the Furthest Point Sampling (FPS) using the evaluated scores to retain more contextual points in downsampling. Then a Geometry-aware Feature Extraction (GeoFE) module is proposed to add geometric information to enrich the awareness of geometric structure in feature aggregation. To evaluate the performances of the proposed methods, we conduct extensive experiments on three difficulty levels of Car class in KITTI dataset. The experimental results show that on "moderate" instances, our results outperform the state-of-the-art method by 1.08%. Moreover, FocusSA is easy to be plugged in popular architectures. [ABSTRACT FROM AUTHOR]

Recently, 3D object detection by point cloud data processing has been applied to robotics and autonomous driving because of the popularity of the LiDAR sensors. Point cloud data contain the depth and geometric space information of an object as compared with the 2D images, and achieve high precision for classification and location. In the traditional processing of point cloud data, the disorder and sparsity of the points are significant problems. In addition, the traditional detector can only support processing a limited number of point clouds. Thus, it is difficult to detect objects using a large number of point clouds. However, the previous methods need to sample the point cloud data into a coarser type, so they cannot avoid the loss of information and the accuracy is affected, as seen in in PV-RCNN. In this paper, we propose a multiscale feature fusion detector called multiscale region proposal networks (MS-RPNs), which can provide multiscale prediction results for difficult category objects. Meanwhile, our method can improve the detection accuracy for smaller objects with the optimal processing of the multiscale feature extraction module. The efficiency and accuracy of the multiscale region proposal network on the KITTI 3D object detection datasets was evaluated using numerous experiments. [ABSTRACT FROM AUTHOR]

The strong demand of autonomous driving in the industry has led to vigorous interest in 3D object detection and resulted in many excellent 3D object detection algorithms. However, the vast majority of algorithms only model single-frame data, ignoring the temporal clue in video sequence. In this work, we propose a new transformer, called Temporal-Channel Transformer (TCTR), to model the temporal-channel domain and spatial-wise relationships for video object detecting from Lidar data. As the special design of this transformer, the information encoded in the encoder is different from that in the decoder. The encoder encodes temporal-channel information of multiple frames while the decoder decodes the spatial-wise information for the current frame in a voxel-wise manner. Specifically, the temporal-channel encoder of the transformer is designed to encode the information of different channels and frames by utilizing the correlation among features from different channels and frames. On the other hand, the spatial decoder of the transformer decodes the information for each location of the current frame. Before conducting the object detection with detection head, a gate mechanism is further deployed for re-calibrating the features of current frame, which filters out the object-irrelevant information by repetitively refining the representation of target frame along with the up-sampling process. Experimental results reveal that TCTR achieves the state-of-the-art performance in grid voxel-based 3D object detection on the nuScenes benchmark. [ABSTRACT FROM AUTHOR]