面向人眼感知的横纵向色差量化评价方法OA
Lateral and Longitudinal Chromatic Aberration Quantitative Evaluation Method for Human Eye Perceptual Sense
紫边与黄绿边是摄影领域一种常见的画质缺陷,常出现在拍摄内容中存在高反差的区域,通常是由于镜头色散或传感器成像面积过小等原因造成的横纵向色差所导致.提出了一种结合Lab色彩空间特点与HSV色彩空间特点的横纵向色差评价与量化方法,相较传统的基于图像RGB三通道的分析方法,该方法与人眼感知更为接近,可以准确判断出横纵向色差所导致的伪像颜色,并计算出准确宽度.实验结果表明,新方法对横纵向色差所产生的伪像颜色的判断准确率达 94.5%,并且相较传统方法,新方法对横纵向色差所产生的伪像宽度计算准确性提升27%.
Purple and yellow-green fringing is a very common image quality flaw in the photography industry.It usually occurs in high-contrast areas of an image.This type of fringing is typically caused by chromatic aberration,which stems from factors such as lens design defects or an excessively small sensor size.A new quantitative evaluation method that combines the advantages of the Lab and HSV color spaces is proposed.Compared with the traditional evaluation method based on the RGB channels of an image,the new method is much closer to human visual perception.It can accurately detect the exact color and width of chromatic aberration artifacts.Experimental results show that the color detection precision of the new method reaches 94.5%,and the detection precision of artifact width is increased by 27%compared with the traditional method.
李湜文;李蔷;周雄飞;任慧娇;潘慧萍;李嘉磊;王鑫怡
北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085北京小米移动软件有限公司,北京 100085
通用工业技术
光学成像测量横向色差纵向色差色边伪像紫边人眼感知图像质量评价
optical imaging measurementlateral chromatic aberrationlongitudinal chromatic aberrationcolor fringing artifactpurple fringinghuman eye perceptualimage quality evaluation
《计量学报》 2026 (3)
334-340,7
评论