激光打印机成像质量综合评价方法研究

发布时间：2019-01-18 11:04

【摘要】：激光打印机凭借低噪声、高质量、高性价比、快速打印等优点,应用领域不断延伸,从传统的商业办公逐步进入到日常家庭。激光打印机成像输出过程中,由于设备输出的精度、机器运转的震动、墨粉转印时存在的磁场引力等,致使打印输出图像不够清晰,呈现噪声和畸变等失真现象。激光打印机的生产过程中,为了减少不合格设备流入市场的几率,需要提前对激光打印机的成像质量进行系统的评价及分析。研究激光打印机成像质量的评价方法对于保障其成像质量及研制生产有着重要的实用意义。目前的激光打印机成像质量评价研究主要是针对线条、网点、字符、图像等不同对象进行独立的质量评价。本文通过分析激光打印机成像质量的客观评价指标,结合主观评价建立了基于支持向量机(Support Vector Machine,SVM)的激光打印机成像质量综合评价模型。首先设计试验样本,通过激光打印机成像,获取线条及字符、实地区域、圆点和图像作为样本。然后,采用不同质量指标评价激光打印机成像质量,并计算试验样本指标值。本文选择以下质量评价指标,包括线条及字符的线宽误差、粗糙度、模糊度和暗度,实地区域的灰度标准偏差、颗粒度和色斑,圆点圆度误差以及图像的结构相似度、清晰度和纹理信息。最后,对样本进行等级分类的主观评价,并利用SVM建立客观质量指标与主观评价结果之间的关系模型,对激光打印机成像质量进行综合评价。本文通过研究分析确定影响激光打印机成像质量的多个关键评价指标,并利用机器学习的方法结合主观评价对质量指标值进行训练分类得到激光打印机成像质量评价模型。试验结果表明,本文提出的关系模型对激光打印机成像质量的综合评价是有效可行的。
[Abstract]:With the advantages of low noise, high quality, high cost performance and fast printing, laser printer has been extended from traditional commercial office to daily home. In the process of laser printer imaging output, due to the accuracy of the output of the equipment, the vibration of the machine operation, and the magnetic field gravity in the transfer printing of toner, the printing output image is not clear enough, and the distortion phenomena such as noise and distortion appear. In order to reduce the probability of substandard equipments flowing into the market, the imaging quality of laser printers should be systematically evaluated and analyzed in advance during the production of laser printers. It is of great practical significance to study the evaluation method of laser printer imaging quality for ensuring its imaging quality and development and production. At present, the research of laser printer imaging quality evaluation is mainly aimed at line, dot, character, image and other different objects for independent quality evaluation. By analyzing the objective evaluation index of laser printer imaging quality and combining subjective evaluation, a comprehensive evaluation model of laser printer imaging quality based on support vector machine (Support Vector Machine,SVM) is established in this paper. First, the experimental samples are designed, and the lines and characters, field areas, dots and images are obtained by laser printer imaging. Then, the laser printer imaging quality is evaluated by different quality indexes, and the test sample index value is calculated. This paper selects the following quality evaluation indicators, including line width error, roughness, ambiguity and darkness of lines and characters, grayscale standard deviation of field area, particle size and color spot, roundness error of dot and structural similarity of image. Clarity and texture information. Finally, the subjective evaluation of the classification of samples is carried out, and the relationship model between objective quality index and subjective evaluation result is established by using SVM to comprehensively evaluate the imaging quality of laser printer. In this paper, several key evaluation indexes affecting the imaging quality of laser printer are determined through research and analysis, and the evaluation model of laser printer imaging quality is obtained by using machine learning method combined with subjective evaluation to train and classify the quality index value. The experimental results show that the proposed relational model is effective and feasible for the comprehensive evaluation of laser printer imaging quality.
【学位授予单位】：武汉大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP334.84

【相似文献】