An improved mean filter algorithm based on gray-scale difference and its application in X-ray detection system

In X-ray detection system, the image may be noised by the interference of various factors such as X-ray scattering, quantum fluctuation etc. For the purpose of not affecting the judgment of defects, it is necessary to remove noise from the image, meanwhile to keep the image edges as clear as before denoising. To achieve this goal, this paper proposed a new method of mean filter algorithm based on gray scale difference. The basic theory is that: Set a threshold value; calculate the absolute value of gray level differences between neighboring pixels and the center pixel. If the absolute value is larger than the threshold, set the weight of the corresponding neighboring pixel to 0. Otherwise set the weight of the pixel to the difference of the threshold and the absolute value. By denoising the image in actual production, this method showed good performance in noise reduction and keeping the image edges clear.