改进YOLOX网络的轴承缺陷小目标检测方法

针对深度学习模型在工业轴承表面缺陷检测中多目标情形下的小目标漏检率高、模型特征融合不充分的问题,基于YOLOX提出一种多注意力特征加权融合的小目标缺陷检测算法。在骨干网络引入特征提取更加细粒度的Res2Block模块,同时嵌入自注意力机制,增加隐性小目标的区域特征,减少漏检率;设计内嵌坐标注意力并作为加权条件的双路金字塔特征融合网络,提升浅层细节特征和深层高级语义特征的交互融合能力;后处理阶段引入Focal Loss损失函数,增加模型对正样本目标的学习,进一步减少漏检率。实验结果表明,与原YOLOX算法相比,改进算法在自制小型列车轴承表面缺陷数据集上mAP提高了4.04个百分点,对小目标的识别率明显提升。     

Analysis on influencing factors of detecting chemical oxygen demand in water by ultraviolet absorption spectroscopy

Taking the standard solution and actual water samples as the research object, this paper studies the influence of various environmental factors (temperature, alkalinity (pH), inorganic salt ions and salinity) on the detection of chemical oxygen demand (COD) in water by ultra-violet (UV) spectrometry, analyzes the influence mechanism, and draws relevant conclusions. The results show that the UV absorbance of the COD standard solution and the actual river water samples varies very little with temperature. The value of pH has certain impact on both the standard solution and the actual water samples, but the pH values of the affected experimental samples are in different ranges. The effects of nitrogen salt ions (NO2- and NO3-) on the UV absorption spectra can be eliminated by intercepting the UV spectrum interval less affected for modeling. The influence of salinity values below 35‰ on the UV absorption spectrum of water is mainly concentrated in the spectral band of 200—250 nm, and the influence above 250 nm is very little. This paper provides a preliminary analysis and discussion on the influence mechanism and laws of various environmental factors in the detection of water quality parameters by UV spectrometry, which provides an experimental basis for selecting the optimal test conditions when establishing the COD prediction model.     

不同预冷温度对甜樱桃果实保鲜效果的影响

目的 探究不同预冷温度对采后甜樱桃果实生理品质的影响,找出最适宜的预冷温度,进而推动甜樱桃产业化发展。方法 本实验分别将采后甜樱桃分别预冷至15、12、9、6、3、0℃后回温至室温,以不进行预冷处理为对照组,比较6种处理方式果实预冷和回温过程中的温度变化,并分别测定回温前后果实的色差、呼吸强度及乙烯释放量、硬度、可溶性固形物(TSS)、风味、感官评价,总酚、类黄酮、花色苷、维生素C(Vc)、可溶性蛋白、丙二醛(MDA)含量以及超氧阴离子产生速率(O₂^-)。结果不同预冷温度均可较好地保持果实外观及风味,能够降低甜樱桃果实呼吸强度与乙烯释放量,抑制果实TSS、MDA含量和O₂^-产生速率的上升,延缓甜樱桃果实硬度的下降,从而保持甜樱桃果实的外观品质,且有利于维持甜樱桃果实中的Vc、可溶性蛋白和花色苷含量,以0℃预冷效果最佳,感官评分最高。结论 预冷能够有效维持甜樱桃果实采后的品质,达到增加甜樱桃果实经济价值的作用。     

基于激光点云数据的叶片型面三维重构

在四坐标叶片型面检测基础上,提出基于激光点云数据的叶片型面三维重构方法。激光位移传感器对叶片型面进行多视角扫描采样,快速采集叶片型面海量点云数据。运用由点到线到面的数学建模原理,先基于端点一阶导矢连续法拟合出光顺NURBS曲线,再依据分片能量法构建辅助曲面拟合出光顺NURBS曲面,最后对分片NURBS曲面进行统一描述,构造出精确光滑的叶片型面,实现叶片型面的三维重构。实验结果表明：采用该方法实现对各类叶片复杂型面的三维重构,重构误差均<0.015 mm,能够满足精密零件的测量需求。     

基于特征值分解的图像边界与空间关系描述

针对图像识别、检索、配准中的特征提取与特征描述问题,利用加权图的邻接矩阵特征值分解方法,构造具有平移、缩放、旋转不变性的图像边界形状和空间关系描述子。理论分析与仿真实验结果表明,该算法在缩放和旋转条件下具有较强的鲁棒性和较好的类别可分离性,相比传统算法可以更好地进行图像描述,且计算量有所减小。     

基于自然接触传感器的智能控制手柄

研究开发了一种用于控制手柄的自然接触传感器组.该传感器组能够实时地测量皮肤血流容积、皮肤导电性、皮肤温度和握力等人体生理信号,同时不会改变和打扰人原来的工作状态.传感器的设计是基于自然接触理念,通过包覆的方式安置到控制手柄上,以保证人和传感器的自然接触,包覆有此类传感器的手柄称为智能控制手柄,实验结果显示智能控制手柄具...     

Boosting k-nearest neighbor classifier by means of input space projection

The k-nearest neighbors classifier is one of the most widely used methods of classification due to several interesting features, such as good generalization and easy implementation. Although simple, it is usually able to match, and even beat, more sophisticated and complex methods. However, no successful method has been reported so far to apply boosting to k-NN. As boosting methods have proved very effective in improving the generalization capabilities of many classification algorithms, proposing an appropriate application of boosting to k-nearest neighbors is of great interest.Ensemble methods rely on the instability of the classifiers to improve their performance, as k-NN is fairly stable with respect to resampling, these methods fail in their attempt to improve the performance of k-NN classifier. On the other hand, k-NN is very sensitive to input selection. In this way, ensembles based on subspace methods are able to improve the performance of single k-NN classifiers. In this paper we make use of the sensitivity of k-NN to input space for developing two methods for boosting k-NN. The two approaches modify the view of the data that each classifier receives so that the accurate classification of difficult instances is favored.The two approaches are compared with the classifier alone and bagging and random subspace methods with a marked and significant improvement of the generalization error. The comparison is performed using a large test set of 45 problems from the UCI Machine Learning Repository. A further study on noise tolerance shows that the proposed methods are less affected by class label noise than the standard methods.     

汉英-泰互译有声语料的数据库研究

“汉英-泰互译有声语料库”的开发,在泰文舆情分析领域,解决了词典分词算法中训练语料缺乏的问题. 本文采用计算机化信息处理技术,对大量的收集来的泰文语料进行整理、规范、加工与存储,统计出泰文词汇8000多个. 然后利用词典翻译和人工校对其进行语料对齐. 最后,结合泰文语言语法特征以及句法的语义特点,分类归纳和规范标注泰语语料,构建了5万条左右的的汉英-泰语料数据库.     

边境市场跨界民族经济交往研究——基于云南地区边境市场的实证分析

跨界民族是中国边境沿线重要的主体构成,其在边境市场中的经济交往对于边境地区的安全、稳定与发展具有重要的作用。边境市场跨界民族的经济交往处于边境市场特殊的“共振”效应之中,包括国家间关系共振、民族间关系共振、市场间关系共振和区域间关系共振。从跨界民族在边境市场上的镜像看,其仍然具有原初市场的痕迹,并开始浸染现代市场带来的变化,此中转变尚需要时间和过程。如何在共振的效应中并建立现代市场的秩序,除了具备客观理性、主观理性外,也需要跨界民族主体间性的交往理性。     

固定测斜仪在面板堆石坝中的应用实例

目前面板堆石坝技术的发展促进监测技术的提高,监测仪器精度越来越高,因此固定式测斜仪以其自身优势在面板堆石坝中的应用范围也在不断扩大,以两个面板堆石坝工程安全监测实例介绍了固定式测斜仪在中国的三种应用即监测防渗墙水平变形、堆石体沉降变形及面板挠度变形,其中监测堆石体沉降变形尚属首次,与相关工程进行了简单对比,为固定式测斜仪在面板堆石坝监测技术发展提出了几点建议。