Scene-based Blind and Flickering Pixel Dynamic Correction Algorithm

As for infrared focal plane array (IRFPA), the correction of bad pixel has always been one of the core issues of infrared non-uniformity correction (NUC). Bad pixels are divided into blind pixels and flickering pixels, while the number of bad pixels in the IRFPA will increase gradually with time, resulting in a deterioration of infrared image quality. Based on the characteristics of bad pixels, similar to the salt and pepper noise in the two-point non-uniformity corrected image, this paper uses the adaptive median filtering (AMF) method to detect real-time blind and flickering pixels of moving and stationary scene. A human-vision-based algorithm is proposed to compensate for the misjudged blind pixels and flickering pixels. The experimental results of the proposed algorithm show that the proposed algorithm has higher robustness to scene changes and better correction effect on blind and flickering pixels than the traditional algorithm.