红外焦平面阵列盲元补偿算法研究

Study on Blind-pixel Compensation Algorithm for Infrared Focal Plane Arrays

提出了一种补偿盲元的新算法,它首先利用自适应中值滤波检测红外焦平面阵列上可能的盲元点,然后再使用细节保持变分法对所挑出的盲元点进行补偿。这样,由于既保护了边缘又抑制了噪声,因此恢复出的图像质量明显优于传统的平滑滤波方法。仿真结果表明,本算法对噪声水平为25%的图像进行恢复后,其峰值信噪比仍可达到34.918 dB。

A novel algorithm is described for compensating blind pixels in infrared focal plane arrays, which is separated into two stages. First, an adaptive median filter is exploited to identify pixels which are likely to be blind. Second, these bad-detector candidates are compensated by using a detail-preserving variational method. Due to edge preservation and noise suppression, the proposed algorithm shows a better performance compared with traditional smoothing filters. Moreover, the experimental results indicate that images with a noise level 25 % can be restored by this method, and the peak signal--to--noise ratio can be up to 34. 918 dB.