微光夜视仪分辨力客观评价判据研究

OBJECTIVE EVALUATION CRITERION FOR THE RESOLUTION OF LOW-LIGHT-LEVEL NIGHT VISION SYSTEMS

本文利用灰度 -梯度共生矩阵模型对微光夜视仪分辨力测量图像的纹理特征进行了分析 ,通过对大量图像的纹理特征参数的计算 ,结果表明 ,梯度熵能对微光夜视仪分辨力测量图像的纹理特征作出有效的客观评价 ,其计算结果与人眼的观察结果具有较好的一致性。通过对人眼观察结果与计算结果比较表明 。

A gray level-gradient cooccurrence matrix model is applied for the texture analysis of an image measuring low-light-level night vision system resolution.The texture parameters are calculated by using this model for a good deal of images. The results show that gradient entropy can be used to evaluate objectively the texture character of an image measuring low-light-level night vision system resolution.Results of calculation are consistent with those from human subjective evaluation. It shows that by comparing the observed and calculated results the gradient entropy can be used as the criterion when measuring objectively the resolution of low-light-level night vision systems.