摘要
基于微扫描的超分辨成像技术在成像系统光学参数不变的基础上提高图像分辨率,常用于红外成像领域。目前,在基于微扫描的红外超分辨成像研究中,采用理想的正方形模型来仿真红外探测器的感光面形状。但是,由于读出电路的存在,实际探测器产品中像素感光面形状并非严格的正方形,使用正方形模型仿真会导致仿真结果和研究结论的不准确。推导了长方形和"折形"两种感光面模型对成像的影响,并通过与正方形模型的仿真对比确定相应的误差,最后结合实验对前述研究成果进行了验证。结果表明:极限情况下,相对于"折形"模型,使用方形模型仿真的误差高达41%;相同像素填充率时,"折形"感光面越远离中心,对成像的MTF影响越大;基于探测器真实感光面形状的仿真模型能有效提高对红外超分辨成像研究的准确性。
Super-resolution (SR) imaging is an effective and economical way to increase the image resolution in spite of the detector limitation, therefore is widely used in many infrared (IR) focal plane array (FPA) imaging application. The ideal square pixel active area (PAA) model is widely used in research of microscanning (MS) based IR SR imaging. However, according to the micrographs of practical IRFPA productions, the PAAs of these productions are actually not a strict square. Three different PAA shapes were analyzed to discuss their effect in MS based IR SR imaging, and experimental results were coincident to theoretical simulation. The results indicate that: using square PAA causes 41%inaccuracy in some limit cases; more closer the "Z shape" to rectangle, higher value of "Z shape" MTF;simulation model based detector active area shape can effectively improve accuracy for IR SP research.
出处
《红外与激光工程》
EI
CSCD
北大核心
2014年第10期3183-3187,共5页
Infrared and Laser Engineering
基金
国家青年自然科学基金(61307021)