无参考图像质量评价的视觉测量照明

为提高视觉测量系统的图像质量,通过分析成像器件和照明光源的光谱匹配特性及照明光源的光谱-色差特性,研究了照明系统的光谱对图像质量的影响;引入对比度信噪比(constrast signal noise ratio,CSNR)和灰度清晰度Tenengrad函数的无参考图像质量的评价方法对不同照明光谱下的图像对比度和清晰度分别进行评价,通过评价函数获得视觉测量照明系统的光谱优化选择原则;最后对相同照度下的单色光和白光(复色光)照明获取的图像的对比度和清晰度成像指标进行对比实验,用所提出的评价函数对不同光谱条件下获取的数字图像进行质量评价.结果表明,使用接近于CCD光谱响应函数峰值波长附近的光谱照明,其图像对比度和清晰度指标都优于用复色光照明所获得图像的指标,为从事视觉测量系统设计者选择照明光源光谱提供了一个参考.

Illumination Optimization for Vision Measurement Based on No-reference Image Quality Assessment

To enhance the image quality of the measuring system based on visual imaging,by analyzing the imaging devices and lighting of the light source spectral imaging system,and matching characteristics of the aberration characteristics,this paper studies the influence on image quality of the spectrum of the lighting system.No-reference image quality evaluation methods,such as constrast signal noise ratio(CSNR) and Tenengrad function,are introduced to analyze the ability of image contrast and clarity under different spectrum lightings.Based on this condition,spectrum optimization principle for imaging,measurement and lighting system is obtained.According to this theory,a spectrum lighting comparison experiment including the monochromatic and the polychromatic light in the same illumination was carried out.The index of quality evaluation of digital image contrast and definition from different spectrum conditions was made by the suggested evaluation function.Result shows that the image quality index of contrast and definition under polychromatic light are inferior to that by using pure spectrum lighting,which is near to the peak wavelength of CCD spectral response function in the experiment.Therefore,it give us an optimization guide for the spectral characteristics when we choose the lighting source system in designing a system based on vision.