摘要
为了解决回扫补偿型红外成像系统图像退化程度在线评价问题,采用了基于红外图像梯度相似度的质量评价方法。在结构相似度算法思想的基础上,针对回扫补偿型红外成像系统的特点,提取出能有效度量图像失真变化程度的梯度特征,计算出各子图像块的梯度相似度均值,由此获得整幅图像的梯度相似度。结果表明,对于退化程度相同的图像,测得结构相似度算法的评价数值从0.772下降到0.705,梯度相似度算法的评价数值由0.700下降到0.543。梯度相似度算法对不同退化程度的红外周扫图像的质量变化比常用的结构相似度算法更为灵敏,且不需要平行光管等外部测试设备,这为回扫补偿型红外成像系统在线图像质量评价及优化提供了一种新的方法。
In order to solve the problem of online assessment of image degradation in retrace compensation infrared imaging system,a quality evaluation method based on the gradient similarity of infrared images was adopted.Based on the idea of a structural similarity algorithm,according to the characteristics of a flyback compensation infrared imaging system,the gradient feature that could effectively measure the degree of image distortion change was extracted,and the average value of the gradient similarity of each sub-image block was calculated,thereby the gradient of the entire image similarity was obtained.The results show that for images with the same degree of degradation,the measured assessment value of the structural similarity algorithm drops from 0.772 to 0.705,and the assessment value of the gradient similarity algorithm drops from 0.700 to 0.543;the gradient similarity algorithm is more sensitive to the quality changes of infrared scan images with different degrees of degradation than the commonly used structural similarity algorithm and does not require external test equipment such as collimators.This study provides a new method for online image quality assessment and optimization of infrared retrace compensation system.
作者
刘大通
刘洋
刘力双
LIU Datong;LIU Yang;LIU Lishuang(School of Instrument Science and Optoelectronics Engineering,Beijing Information Science and Technology University,Beijing 100192,China)
出处
《激光技术》
CSCD
北大核心
2024年第1期121-126,共6页
Laser Technology
关键词
成像系统
图像质量评价
结构相似度
梯度相似度
imaging systems
image quality assessment
struetural similarity
gradient similarity
