OCR(Optical Character Recognition)是通过检测字符每个像素亮度的模式确定其形状,然后用字符识别方法将形状翻译成计算机文字的过程。文章利用Java语言实现OCR步骤,包括像素二值化,图像分割,训练识别和输出等。测试开发是在web验证码...OCR(Optical Character Recognition)是通过检测字符每个像素亮度的模式确定其形状,然后用字符识别方法将形状翻译成计算机文字的过程。文章利用Java语言实现OCR步骤,包括像素二值化,图像分割,训练识别和输出等。测试开发是在web验证码识别场景中进行的,web验证码是将一串随机产生的符号,生成为图片,再加上一些干扰线,使之能有效防止恶意注册和灌水。通过测试表明,该方法可行、有效;拒识率、误识率低;识别速度快,具有一定的实用意义。展开更多
电能表的检定误差是评价电能表质量的重要指标之一。当前,行业内主要以基本误差限作为合格与不合格的评价指标,缺乏对合格指标的进一步细分。此外,人工处理的工作量巨大。针对这些问题,提出了一种基于两阶段聚类的电能表质量评价模型。...电能表的检定误差是评价电能表质量的重要指标之一。当前,行业内主要以基本误差限作为合格与不合格的评价指标,缺乏对合格指标的进一步细分。此外,人工处理的工作量巨大。针对这些问题,提出了一种基于两阶段聚类的电能表质量评价模型。首先分析了各种聚类方法的适用范围与A、B、C三个厂家单相电能表的检定数据,选择K-means算法与HDBSCAN(Hierarchical Density-Based Spatial Clustering of Applications with Noise)作为分析方法。随后,在第一阶段聚类采用HDBSCAN确定K-means算法的K值和初始聚类中心,以解决初始聚类中心随机化的问题。在第二阶段聚类采用K-means算法对所有三个厂家的检定数据进行聚类,得出各聚类类别的统计量描述以确定类别的优劣性,并对结果进行验证。最后,计算得出各厂家和各型号的检定数据在不同类别的占比,给出评价模型的一般推广。所提方法可为电网公司单相电能表厂家与型号的选择提供参考。展开更多
Error modelling and compensating technology is an effective method to improve the processing precision.The position and orientation deviation of workpiece is caused by the fixing and manufacturing errors of the fixtur...Error modelling and compensating technology is an effective method to improve the processing precision.The position and orientation deviation of workpiece is caused by the fixing and manufacturing errors of the fixture.How to reduce the position and orientation deviation of workpiece has become a technical problem of improving the processing quality of workpiece.In order to increase machining accuracy,an implementation scheme of fixture system comprehensive errors（FSCE） compensation is proposed.A FSCE parameter model is established by analyzing the influence of contact points on the position and orientation of workpiece.Meanwhile,a parameter identification method for FSCE parameter model is presented by using the 3-2-1 deterministic positioning fixture,which determines the model parameters.Moreover,a FSCE compensation model is formulated to study the compensation value of the cutting position.By using RenishawOMP60 Probe and combining vertical machining centre（SKVH850） equipment with SKY2001 Open CNC System,on-machine verification system（OMVS） is built to measure FSCE successfully.The processing error can be reduced by analyzing the cutting position of the tool with the homogeneous transformation of space coordinate system.Finally,the compensation experiment of real time errors is conducted,and the cylindricality and perpendicularity errors of hole surface are reduced by 30.77% and 28.57%,respectively.This paper provides a new way of realizing the compensation of FCSE,which can improve the machining accuracy of workpiece largely.展开更多
文摘OCR(Optical Character Recognition)是通过检测字符每个像素亮度的模式确定其形状,然后用字符识别方法将形状翻译成计算机文字的过程。文章利用Java语言实现OCR步骤,包括像素二值化,图像分割,训练识别和输出等。测试开发是在web验证码识别场景中进行的,web验证码是将一串随机产生的符号,生成为图片,再加上一些干扰线,使之能有效防止恶意注册和灌水。通过测试表明,该方法可行、有效;拒识率、误识率低;识别速度快,具有一定的实用意义。
文摘电能表的检定误差是评价电能表质量的重要指标之一。当前,行业内主要以基本误差限作为合格与不合格的评价指标,缺乏对合格指标的进一步细分。此外,人工处理的工作量巨大。针对这些问题,提出了一种基于两阶段聚类的电能表质量评价模型。首先分析了各种聚类方法的适用范围与A、B、C三个厂家单相电能表的检定数据,选择K-means算法与HDBSCAN(Hierarchical Density-Based Spatial Clustering of Applications with Noise)作为分析方法。随后,在第一阶段聚类采用HDBSCAN确定K-means算法的K值和初始聚类中心,以解决初始聚类中心随机化的问题。在第二阶段聚类采用K-means算法对所有三个厂家的检定数据进行聚类,得出各聚类类别的统计量描述以确定类别的优劣性,并对结果进行验证。最后,计算得出各厂家和各型号的检定数据在不同类别的占比,给出评价模型的一般推广。所提方法可为电网公司单相电能表厂家与型号的选择提供参考。
基金supported by National Natural Science Foundation of China (Grant No. 50975200)National Key Technologies R & D Programmer of China (Grant No. 2009ZX04014-021)
文摘Error modelling and compensating technology is an effective method to improve the processing precision.The position and orientation deviation of workpiece is caused by the fixing and manufacturing errors of the fixture.How to reduce the position and orientation deviation of workpiece has become a technical problem of improving the processing quality of workpiece.In order to increase machining accuracy,an implementation scheme of fixture system comprehensive errors（FSCE） compensation is proposed.A FSCE parameter model is established by analyzing the influence of contact points on the position and orientation of workpiece.Meanwhile,a parameter identification method for FSCE parameter model is presented by using the 3-2-1 deterministic positioning fixture,which determines the model parameters.Moreover,a FSCE compensation model is formulated to study the compensation value of the cutting position.By using RenishawOMP60 Probe and combining vertical machining centre（SKVH850） equipment with SKY2001 Open CNC System,on-machine verification system（OMVS） is built to measure FSCE successfully.The processing error can be reduced by analyzing the cutting position of the tool with the homogeneous transformation of space coordinate system.Finally,the compensation experiment of real time errors is conducted,and the cylindricality and perpendicularity errors of hole surface are reduced by 30.77% and 28.57%,respectively.This paper provides a new way of realizing the compensation of FCSE,which can improve the machining accuracy of workpiece largely.
