MFF-Net: Multimodal Feature Fusion Network for 3D Object Detection

In complex traffic environment scenarios, it is very important for autonomous vehicles to accurately perceive the dynamic information of other vehicles around the vehicle in advance. The accuracy of 3D object detection will be affected by problems such as illumination changes, object occlusion, and object detection distance. To this purpose, we face these challenges by proposing a multimodal feature fusion network for 3D object detection (MFF-Net). In this research, this paper first uses the spatial transformation projection algorithm to map the image features into the feature space, so that the image features are in the same spatial dimension when fused with the point cloud features. Then, feature channel weighting is performed using an adaptive expression augmentation fusion network to enhance important network features, suppress useless features, and increase the directionality of the network to features. Finally, this paper increases the probability of false detection and missed detection in the non-maximum suppression algorithm by increasing the one-dimensional threshold. So far, this paper has constructed a complete 3D target detection network based on multimodal feature fusion. The experimental results show that the proposed achieves an average accuracy of 82.60% on the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) dataset, outperforming previous state-of-the-art multimodal fusion networks. In Easy, Moderate, and hard evaluation indicators, the accuracy rate of this paper reaches 90.96%, 81.46%, and 75.39%. This shows that the MFF-Net network has good performance in 3D object detection.     

基于多尺度融合方法的无人机对地车辆目标检测算法研究

由于飞行高度等原因,无人机图像在实际使用中目标尺寸普遍较小、特征信息不明显,使用现有的算法对其进行目标检测存在困难。因此,提出了基于多尺度融合的图像多目标检测方法,使用Faster R-CNN为基础框架,将不同层次的特征信息进行融合,再结合上下文信息,实现了对无人机图像小目标检测。使用VisDrone2019数据集对地面车辆进行目标检查,实验证明:无人机对地面车辆目标的检测达到了较好的结果,所使用算法的精度达到88%,与其它算法相比提升了3.8%以上。     

基于改进YOLOv4算法的轮毂表面缺陷检测

汽车轮毂加工过程中产生的表面缺陷严重影响整车的美观性及服役性能,针对人工检测效率低、漏检率高的问题,提出一种基于改进YOLOv4算法的轮毂表面缺陷检测方法。构建了轮毂缺陷数据集,其包含6种表面缺陷,由2346张4928×3264pixel的图像组成;采用K-means方法进行先验框聚类,并针对YOLOv4算法在纤维、粘铝等小尺度缺陷上检测精度不足问题,在原网络Neck部分引入细化U型网络模块(TUM)和注意力机制,用于增强有效特征并抑制无效特征,强化多尺度特征提取与融合,改善特征处理过程中可能存在的小目标信息丢失问题;基于该数据集,训练并测试不同算法的缺陷检测性能并验证改进模块的有效性。结果表明,该方法大幅提升了粘铝等小尺寸缺陷的检测能力,缺陷检测平均精度达到85.8%,与多种算法相比较检测精度最高。     

Unsupervised Domain Adaptation Based on Discriminative Subspace Learning for Cross-Project Defect Prediction

Cross-project defect prediction (CPDP) aims to predict the defects on target project by using a prediction model built on source projects. The main problem in CPDP is the huge distribution gap between the source project and the target project, which prevents the prediction model from performing well. Most existing methods overlook the class discrimination of the learned features. Seeking an effective transferable model from the source project to the target project for CPDP is challenging. In this paper, we propose an unsupervised domain adaptation based on the discriminative subspace learning (DSL) approach for CPDP. DSL treats the data from two projects as being from two domains and maps the data into a common feature space. It employs cross-domain alignment with discriminative information from different projects to reduce the distribution difference of the data between different projects and incorporates the class discriminative information. Specifically, DSL first utilizes subspace learning based domain adaptation to reduce the distribution gap of data between different projects. Then, it makes full use of the class label information of the source project and transfers the discrimination ability of the source project to the target project in the common space. Comprehensive experiments on five projects verify that DSL can build an effective prediction model and improve the performance over the related competing methods by at least 7.10% and 11.08% in terms of G-measure and AUC.     

一种用于跨域轴承故障诊断的深度自适应网络EI北大核心CSCD

针对轴承在不同工况条件下的振动数据存在分布差异,导致诊断精度低的问题,提出一种新的深度自适应网络用于跨域条件下的轴承故障诊断。采用傅里叶变换将原始时域振动信号转换为频域信号并通过深度特征提取器提取其分类特征;利用最大均值差异(maximize mean discrepancy,MMD)来进行深度特征的边缘分布对齐;利用Wasserstein度量网络将源域中有标签数据的类别结构与目标域中无标签数据的类别结构进行匹配,即对齐不同域的类别条件分布,使得故障数据在不同域的分布能够更好的对齐,从而提高模型在目标域未标签数据集上的分类准确率。试验利用凯斯西储大学公开的故障轴承数据集进行了两种跨域条件的模型迁移,验证了该网络在不同迁移场景中都具有较高的准确率,且优于其他深度自适应网络。     

基于深度学习的合作目标靶球检测

为了解决复杂场景下激光跟踪仪对合作目标靶球的精确识别难题,提出了基于深度学习的合作目标靶球高效检测方法。首先分析了合作目标靶球的图像特征,然后采用改进的YOLOv2模型,针对合作目标靶球多尺度与小目标占比多的特点,提出了一种基于注意力机制的改进方法,同时为提高网络模型对复杂背景的抗干扰能力,提出了一种数据增强方法。测试结果表明,所提出的基于注意力机制与数据增强的改进YOLOv2模型对复杂背景的抗干扰能力较强,且对合作目标靶球的检测精度有显著提高,在合作目标靶球测试集上的检测准确率达到92.25%,能够有效满足激光跟踪仪在大型装置精密装配过程中的目标检测精度需求。     

深度领域自适应及其在跨工况故障诊断中的应用

实际生产中,机械设备的工况变化会造成监测数据的分布差异,破坏分类模型的应用基础,降低诊断准确率。为此,提出一种基于深度学习的领域自适应方法,用于跨工况情境下轴承故障诊断。该方法构建两个级联的深度网络:前者用于处理振动信号,自动挖掘故障敏感特征;后者用于将不同工况的样本特征同步映射到一个深度隐藏层(公共特征空间)中,消除工况波动引起的分布差异,生成工况不变特征,实现领域自适应。此外,该深度映射网络可通过参数优化方法自适应构建,能够实现最佳的跨域诊断性能。实验表明,与其他方法和相关研究相比,深度领域自适应在跨工况故障识别中具有更高的准确率。     

MRMR Based Feature Vector Design for Efficient Citrus Disease Detection

In recent times, the images and videos have emerged as one of the most important information source depicting the real time scenarios. Digital images nowadays serve as input for many applications and replacing the manual methods due to their capabilities of 3D scene representation in 2D plane. The capabilities of digital images along with utilization of machine learning methodologies are showing promising accuracies in many applications of prediction and pattern recognition. One of the application fields pertains to detection of diseases occurring in the plants, which are destroying the widespread fields. Traditionally the disease detection process was done by a domain expert using manual examination and laboratory tests. This is a tedious and time consuming process and does not suffice the accuracy levels. This creates a room for the research in developing automation based methods where the images captured through sensors and cameras will be used for detection of disease and control its spreading. The digital images captured from the field's forms the dataset which trains the machine learning models to predict the nature of the disease. The accuracy of these models is greatly affected by the amount of noise and ailments present in the input images, appropriate segmentation methodology, feature vector development and the choice of machine learning algorithm. To ensure the high rated performance of the designed system the research is moving in a direction to fine tune each and every stage separately considering their dependencies on subsequent stages. Therefore the most optimum solution can be obtained by considering the image processing methodologies for improving the quality of image and then applying statistical methods for feature extraction and selection. The training vector thus developed is capable of presenting the relationship between the feature values and the target class. In this article, a highly accurate system model for detecting the diseases occurring in citrus fruits using a hybrid feature development approach is proposed. The overall improvement in terms of accuracy is measured and depicted.     

Fuzzy logic control of vehicle suspensions with dry friction nonlinearity

We design and investigate the performance of fuzzy logic-controlled (FLC) active suspensions on a nonlinear vehicle model with four degrees of freedom, without causing any degeneration in suspension working limits. Force actuators were mounted parallel to the suspensions. In this new approach, linear combinations of the vertical velocities of the suspension ends and accelerations of the points of connection of the suspension to the body have been used as input variables. The study clearly demonstrates the effectiveness of the fuzzy logic controller for active suspension systems. Suspension working space degeneration is the most important problem in various applications. Decreasing the amplitudes of vehicle body vibrations improves ride comfort. Body bounce and pitch motion of the vehicle are presented both in time domain when travelling over a ramp-step road profile and in frequency domain. The results are compared with those of uncontrolled systems. At the end of this study, the performance and the advantage of the suggested approach and the improvement in ride comfort are discussed.     

An efficient and automatic glioblastoma brain tumor detection using shift‐invariant shearlet transform and neural networks

The detection and segmentation of tumor region in brain image is a critical task due to the similarity between abnormal and normal region. In this article, a computer‐aided automatic detection and segmentation of brain tumor is proposed. The proposed system consists of enhancement, transformation, feature extraction, and classification. The shift‐invariant shearlet transform (SIST) is used to enhance the brain image. Further, nonsubsampled contourlet transform (NSCT) is used as multiresolution transform which transforms the spatial domain enhanced image into multiresolution image. The texture features from grey level co‐occurrence matrix (GLCM), Gabor, and discrete wavelet transform (DWT) are extracted with the approximate subband of the NSCT transformed image. These extracted features are trained and classified into either normal or glioblastoma brain image using feed forward back propagation neural networks. Further, K‐means clustering algorithm is used to segment the tumor region in classified glioblastoma brain image. The proposed method achieves 89.7% of sensitivity, 99.9% of specificity, and 99.8% of accuracy.     

基于声呐历程累积图像的弱回波目标检测方法

针对浅海随机噪声与混响背景下蛙人等弱回波强度、慢速小目标的检测问题,提出一种基于声呐历程累积图像的目标检测方法。首先根据声呐图像时域、空域相关性,采用背景空时归一化处理技术,抑制声呐背景中的静态混响、突发性噪声等强回波干扰。声呐历程累积图像集成了多帧声呐图像的信息,目标回波亮点由于运动连续性形成亮线特征,利用该特征,采用Radon恒虚警率(Radon Constant False Alarm Rate,Radon-CFAR)检测声呐历程累积图像中的目标短时运动轨迹,能够检测到低信噪比的目标。分析了空时归一化处理和检测算法的性能,并通过海试数据验证了该算法的有效性,可以检测到低信噪比的蛙人目标回波。     

Fire Detection Method Based on Improved Fruit Fly Optimization-Based SVM

Aiming at the defects of the traditional fire detection methods, which are caused by false positives and false negatives in large space buildings, a fire identification detection method based on video images is proposed. The algorithm first uses the hybrid Gaussian background modeling method and the RGB color model to perform fire prejudgment on the video image, which can eliminate most non-fire interferences. Secondly, the traditional regional growth algorithm is improved and the fire image segmentation effect is effectively improved. Then, based on the segmented image, the dynamic and static features of the fire flame are further analyzed and extracted in the area of the suspected fire flame. Finally, the dynamic features of the extracted fire flame images were fused and classified by improved fruit fly optimization support vector machine, and the recognition results were obtained. The video-based fire detection method proposed in this paper greatly improves the accuracy of fire detection and is suitable for fire detection and identification in large space scenarios.     

基于声呐历程累积图像的弱回波目标检测方法

针对浅海随机噪声与混响背景下蛙人等弱回波强度、慢速小目标的检测问题,提出一种基于声呐历程累积图像的目标检测方法.首先根据声呐图像时域、空域相关性,采用背景空时归一化处理技术,抑制声呐背景中的静态混响、突发性噪声等强回波干扰.声呐历程累积图像集成了多帧声呐图像的信息,目标回波亮点由于运动连续性形成亮线特征,利用该特征,采...     

A content-based image retrieval system based on object-moment feature

The last decade has witnessed great interest in research on content-based image retrieval (CBIR). In 2009, Lin et al. proposed a smart CBIR system based on colour and texture feature. Their system has a high detection rate except the cases where image objects have similar shapes. To enhance the detection rate a shape-based image feature called object-moment is proposed in this paper. Object-moment uses the moment of force to compute the object edge feature by calculating the distance from each edge pixel to the axis, and adding them up as a feature. Besides, we integrate the colour features (NSOM, CSOM) and the texture features (CCM, DBPSP) to enhance image detection rate and simplify computation of image retrieval. A series of analyses and comparisons are performed in our experiments to demonstrate that our proposed method improves the retrieval accuracy significantly.     

基于改进YOLOv5算法的PCB裸板缺陷检测

目的 将基于深度学习的YOLOv5算法应用于PCB裸板的缺陷检测上,以提高检测的准确率。方法 通过增加特征融合通路,将C2、C3、C4层直接与P2、P3、P4层相连,从而减小信息的损耗;引入更浅层的C2、F2、P2特征图以增加图像的细节信息;并且使用注意力机制SE_block,大幅提高原算法的准确率。结果 改进后的网络的平均精度由91.54%提高至97.36%,提高了5.82%,并且对于各类缺陷,算法的检测精度都能保持在90%以上,满足工业的需求。结论 文中的算法提高了检测精度,体现了浅层信息在小目标检测上的作用,验证了多信息融合通路的优势,彰显了注意力机制的优越性,相比于原算法具有一定的优势。     

Real-Time Detection of Cracks on Concrete Bridge Decks Using Deep Learning in the Frequency Domain

This paper presents a vision-based crack detection approach for concrete bridge decks using an integrated one-dimensional convolutional neural network (1D-CNN) and long short-term memory (LSTM) method in the image frequency domain. The so-called 1D-CNN-LSTM algorithm is trained using thousands of images of cracked and non-cracked concrete bridge decks. In order to improve the training efficiency, images are first transformed into the frequency domain during a preprocessing phase. The algorithm is then calibrated using the flattened frequency data. LSTM is used to improve the performance of the developed network for long sequence data. The accuracy of the developed model is 99.05%, 98.9%, and 99.25%, respectively, for training, validation, and testing data. An implementation framework is further developed for future application of the trained model for large-scale images. The proposed 1D-CNN-LSTM method exhibits superior performance in comparison with existing deep learning methods in terms of accuracy and computation time. The fast implementation of the 1D-CNN-LSTM algorithm makes it a promising tool for real-time crack detection.     

基于深度学习的烟包识别与分类

目的 提取烟包图像数据训练深度学习目标检测模型,提升烟包流水线拣包效率和准确性。方法 基于深度学习建立一种烟包识别分类模型,对原始YOLOv3模型进行改进,在原网络中加入设计的多空间金字塔池化结构(M–SPP),将64×64尺度的特征图下采样与32×32尺度的特征图进行拼接,并去除16×16尺度的预测特征层,提高模型的检测准确率和速度,并采用K–means++算法对先验框参数进行优化。结果 实验表明该目标检测模型平均准确率达到99.68%,检测速度达到70.82帧/s。结论 基于深度学习建立的图像识别分类模型准确率高且检测速度快,有效满足烟包流水线自动化实时检测。     

车辆交通事故视频监测算法研究

为快速准确地检测出车辆交通事故的发生,本文提出一种基于计算机视觉和图像处理技术的事故监测算法．统计发现车辆交通事故中会因碰撞伴随产生烟雾、碎片,同时车辆的运动速度会发生锐减．故通过提取交通监控视频中的烟雾、碎片和运动目标速度的变化等参数,结合麦夸特法进行公式拟合可得到交通事故发生概率公式,并计算出事故发生概率．实验结果表明,这种方法能有效地监测出车辆交通事故,检测准确率达89．4％,验证了算法的有效性．此方法应用在智能交通系统中将能及时监测到事故发生,可为后续事故处理赢得时间,故有一定的实用价值．     

基于背景抑制与改进多尺度LSD的声呐小目标检测算法

针对声呐小目标检测由于水下环境复杂、目标回波信号弱等因素造成虚警率和误检率较高的问题,文章提出基于背景抑制和改进直线分割检测(Line Segment Detection, LSD)的检测算法。首先对原始声呐数据截取序列片段,构建多周期累积历程图,凸显运动目标轨迹线特征;其次设计边缘滤波算子,有效滤除部分背景噪声,并结合投影变换进行线特征增强,不仅实现了断裂直线重连,还抑制了剩余噪声;然后基于图像金字塔改进了多尺度LSD直线分割检测算法,有效缓解了过检测问题,大幅增加了直线平均长度;最后为了合并冗余检测信息,利用运动轨迹时空一致性特征设计后处理模块,提高了检测定位精度。通过多组无人遥控潜水器(Remotely Operated Vehicle, ROV)、潜水员、空心球靶小目标序列的湖试、海试数据的定量与可视化结果定性分析,实验结果显示,文中算法与传统LSD相比,误检率和漏检率分别降低了11.2和3.9个百分点,定位误差下降了1.495个像素。结果表明,文中所提算法大幅提高了声呐小目标检测精度,为后续水下目标识别、跟踪等任务奠定重要基础。     

A Two-Stage Vehicle Type Recognition Method Combining the Most Effective Gabor Features

Vehicle type recognition (VTR) is an important research topic due to its significance in intelligent transportation systems. However, recognizing vehicle type on the real-world images is challenging due to the illumination change, partial occlusion under real traffic environment. These difficulties limit the performance of current stateof-art methods, which are typically based on single-stage classification without considering feature availability. To address such difficulties, this paper proposes a twostage vehicle type recognition method combining the most effective Gabor features. The first stage leverages edge features to classify vehicles by size into big or small via a similarity k-nearest neighbor classifier (SKNNC). Further the more specific vehicle type such as bus, truck, sedan or van is recognized by the second stage classification, which leverages the most effective Gabor features extracted by a set of Gabor wavelet kernels on the partitioned key patches via a kernel sparse representation-based classifier (KSRC). A verification and correction step based on minimum residual analysis is proposed to enhance the reliability of the VTR. To improve VTR efficiency, the most effective Gabor features are selected through gray relational analysis that leverages the correlation between Gabor feature image and the original image. Experimental results demonstrate that the proposed method not only improves the accuracy of VTR but also enhances the recognition robustness to illumination change and partial occlusion.