非均匀性是制约红外系统成像质量的关键因素之一。提出了非均匀性对目标检测影响程度的分析方法，给出了不同性能的目标检测系统对于非均匀性的指标要求。该方法首先采用点光源成像实验仿真了非均匀性和信杂比的关系，然后根据N-P准则（Neyman-Person Criterion）从理论上推导了信杂比与检测概率和虚警概率的关系。实验结果表明，非均匀性每增大1%，检测概率降低15%左右，当非均匀性达到10%时基本失去目标检测功能。当虚警概率降低到5%左右后将不再受非均匀性的影响。

Non-uniformity is the main constraint for the image quality of infrared systems. Analysis approach for non-uniformity influence on target detection was proposed, and the indicators of non-uniformity were provided to satisfy different target detector systems. First, point light source was used for simulating the relationship between non-uniformity and SCR. Then, the impacts of SCR on probability of detection and false alarm were derived based on Neyman-Person Criterion. Experimental results show that each increase of 1% of non-uniformity will cause probability of detection decrease almost 15%. While non-uniformity reaches 10%, the detection system is basically useless. The probability of false alarm reduces with the increases of non-uniformity, and the non-uniformity has no impact on the probability of false alarm when it decreases to 5%.