基于NETD的红外热像仪图像预处理方法研究

Research on Image Preprocessing Method of Infrared Camera Based on NETD

噪声等效温差(NETD)是红外热像仪的重要参数,它能够体现红外探测器的噪声特性。在剔除无效像元的基础上,利用线性滤波、非线性滤波、直方图均衡化和多尺度图像增强等算法对热像仪原始数据进行图像预处理;用实验者对画质和图像细节的直接观察作为定性判据,噪声等效温差作为定量判据,综合评价图像预处理算法的优劣。通过实验发现,线性滤波后进行多尺度图像增强的方式能够使图像细节更清晰,同时噪声等效温差减小约25.9%,是目前较为适合的实验用热像仪的图像预处理方式。

Noise equivalent temperature difference(NETD)is an important parameter of infrared thermal imager,which can reflect the noise characteristics of the infrared detector.On the basis of eliminating invalid pixels,the image of the thermal imager raw data is preprocessed by linear filtering,nonlinear filtering,histogram equalization and multi-scale image enhancement.Using the direct observation of the image quality and image details by the experimenter as the qualitative criterion and the noise equivalent temperature difference as the quantitative criterion to comprehensively evaluate the advantages and disadvantages of the image preprocessing algorithm.It is found through experiments that the multi-scale image enhancement method after linear filtering can make the image details clearer,and the noise equivalent temperature difference is reduced by about 25.9%,which is the most suitable image preprocessing method for experimental infrared camera.