基于机器视觉的航空铆钉尺寸检测技术

航空铆钉的尺寸偏差会增加航空事故发生的可能性。为了实现航空铆钉尺寸检测的高精度与智能化,提出一种基于机器视觉的航空铆钉尺寸检测方法。通过边缘检测算法对铆钉进行自动识别与检测,可减少航空铆钉尺寸的误差。基于航空铆钉产品规范,分析得出各项尺寸参数要求与事件流程,使用OpenCV作为开发平台,采用阈值分割算法与分水岭算法进行图像分割;同时,通过平滑处理、形态学、轮廓提取算法完成图像处理。实测结果表明,在±0.005mm精度下,铆钉尺寸检测结果与实际值的吻合度达到90%以上。     

SCMA场景下的极化码编解码方案改进研究

极化码是目前唯一一种被证明可达到信道容量的编码方式,稀疏码分多址接入(Sparse Code Division Multiple Access,SCMA)可以提高频谱资源的利用率和接入系统的用户接入数量。为了提升SCMA与极化码的联合系统的误码率性能和译码时延,使联合系统的应用越来越广泛,提出了2种降低译码复杂度的方式:简化的左信息更新方式和剪枝译码算法。简化的左信息更新方式对于N=256,N=1024的极化码分别能降低37.6%和44.6%的存储资源占用数;剪枝译码算法在码率为0.5时能降低50%左右的计算复杂度。基于简化的SCAN算法改进了联合检测译码算法,在接收机采用外循环迭代的结构,引入了阻尼机制,选取最优的阻尼方式和最优阻尼值。仿真分析了所提联合检测译码算法与原有联合检测译码算法的误码率性能差异,外迭代接收机性能与内外双循环迭代的接收机性能相同,采用阻尼机制联合接收机的性能比无阻尼接收机的性能高0.8 dB左右,SJIDD的误帧率性能比保留宽度为32的SCLJDD的性能低0.7 dB左右,但能使接收机处理时延降至原来的1/4~1/8。     

多共振分量融合CNN的行星齿轮箱故障诊断

针对行星齿轮箱中各部件所激起的振动成分混叠、早期故障特征经常被较强的各级齿轮谐波成分以及环境噪声所湮没的问题，提出一种多共振分量融合卷积神经网络（multi-resonance component fusion based convolutional neural network，简称MRCF-CNN）的行星齿轮箱故障诊断方法。首先，对振动信号进行共振稀疏分解，得到包含齿轮谐波成分的高共振分量和可能包含轴承故障冲击成分的低共振分量；其次，构建多共振分量融合卷积神经网络，将得到的高、低共振分量和原始振动信号进行自适应的特征级融合，通过有监督的方式训练模型并进行行星齿轮箱故障诊断。对行星齿轮箱实验数据的分析结果表明，该方法能够有效分类行星齿轮箱中滚动轴承和齿轮的故障，成功对行星齿轮箱故障进行诊断，同时能够进一步增强卷积神经网络对振动信号所蕴含的故障信息的辨识能力。     

Experience-Based Cognition for Driving Behavioral Fingerprint Extraction

Abstract

With the rapid progress of information technologies, cars have been made increasingly intelligent. This allows cars to act as cognitive agents, i.e., to acquire knowledge and understanding of the driving habits and behavioral characteristics of drivers (i.e., driving behavioral fingerprint) through experience. Such knowledge can be then reused to facilitate the interaction between a car and its driver, and to develop better and safer car controls. In this paper, we propose a novel approach to extract the driver’s driving behavioral fingerprints based on our conceptual framework Experience-Oriented Intelligent Things (EOIT). EOIT is a learning system that has the potential to enable Internet of Cognitive Things (IoCT) where knowledge can be extracted from experience, stored, evolved, shared, and reused aiming for cognition and thus intelligent functionality of things. By catching driving data, this approach helps cars to collect the driver’s pedal and steering operations and store them as experience; eventually, it uses obtained experience for the driver’s driving behavioral fingerprint extraction. The initial experimental implementation is presented in the paper to demonstrate our idea, and the test results show that it outperforms the Deep Learning approaches (i.e., deep fully connected neural networks and recurrent neural networks/Long Short-Term Memory networks).     

Self-learning residual model for fast intra CU size decision in 3D-HEVC

As an extension of the High Efficiency Video Coding (HEVC) standard, 3D-HEVC requires to encode multiple texture views and depth maps, which inherits the same quad-tree coding structure as HEVC. Due to the distinct properties of texture views and depth maps, existing fast intra prediction approaches were presented for the coding of texture views and depth maps, respectively. To further reduce the coding complexity of 3D-HEVC, a self-learning residual model-based fast coding unit (CU) size decision approach is proposed for the intra coding of both texture views and depth maps. Residual signal, which is defined as the difference between the original luminance pixel and the optimal prediction luminance pixel, is firstly extracted from each CU. Since residue signal is strongly correlated with the optimal CU partition, it is used as the feature of each CU. Then, a self-learning residual model is established by intra feature learning, which iteratively learns the features of the previously encoded coding tree unit (CTU) generated by itself. Finally, a binary classifier is developed with the self-learning residual model to early terminate CU size decision of both texture views and depth maps. Experimental results show the proposed fast intra CU size decision approach achieves 33.3% and 49.3% encoding time reduction on average for texture views and depth maps with negligible loss of overall video quality, respectively.     

基于差动共焦显微技术的微区拉曼光学系统构建与实验研究

研制了一种基于差动共焦显微技术的微区拉曼光学系统装置,对无机样品进行微区拉曼光谱探测。传统显微共焦拉曼光谱技术没有强调系统的定焦能力,而所研制的光学系统装置利用差动共焦曲线过零点与焦点位置精确对应的特性,采用反馈控制技术,具有长时间定焦功能。在微区拉曼散射信号收集时,采用多模光纤空间耦合技术,以光纤代替传统物理探测针孔,提高了环境抗干扰能力,优化了系统结构和装调性能。实验结果表明:该装置具有较高稳定性,可有效探测单壁碳纳米管在1581.510 cm^-1,2708.065 cm^-1特征峰处的拉曼频移及纯物质硫在153.113 cm^-1,219.917 cm^-1,473.322 cm^-1特征峰处的拉曼频移并且实现碳管的单线检测,满足了光谱探测系统装置设计需求。     

基于自回归谱外推的全聚焦成像分辨力提升

针对相邻缺陷全聚焦超声成像混叠问题，结合低阶、宽有效频带自回归谱外推方法，压缩超声波时域脉冲宽度，实现亚波长级全聚焦（Total focusing method，TFM）成像分辨力。建立碳钢试块模型，设置两个中心间距1.8 mm，直径1.3 mm圆孔，选用中心频率2.25 MHz，32阵元相控阵探头采集全矩阵数据。针对全矩阵数据，选择自回归阶数为2，信号频谱最大幅值下降14 dB为有效频带，建立自回归模型并外推有效频带外的高频与低频成分，随后对全矩阵数据进行延迟叠加处理和TFM成像。仿真结果表明，低阶、宽有效频带自回归谱外推处理方法具有较高的鲁棒性和准确性，TFM成像后可有效分离中心间距0.7λ（λ为超声波长）圆孔，保留缺陷横向位置信息的同时，定位误差不超过0.73%。对碳钢试块中相同位置及尺寸的圆孔进行试验验证，定位误差不超过1.06%，有效地提高TFM成像分辨力。     

Hybrid fully convolutional networks-based skin lesion segmentation and melanoma detection using deep feature

Fully convolutional networks (FCNs) take the input of arbitrary size and produce correspondingly sized output with efficient inference and learning. The automatic diagnosis of melanoma is very essential for reducing the mortality rate by identifying the disease in earlier stages. A two-stage framework is used for implementing the melanoma detection, segmentation of skin lesion, and identification of melanoma lesions. Two FCNs based on VGG-16 and GoogLeNet are incorporated for improving the segmentation accuracy. A hybrid framework is used for incorporating these two FCNs. The classification is done by extracting the feature from segmented lesion by using deep residual network and a hand-crafted feature. Classification is done by support vector machine. The performance analysis of our framework gives a promising accuracy, that is, 0.8892 for classification in ISBI 2016 dataset and 0.853 for ISIC 2017 dataset.