The hybrid median filter is a non-linear class window filter, which easily eliminates impulse noise while retaining the edges. Compared to the basic version of the hybrid filter medium has better features of corners preservation. The basic idea behind the filter is that any element of the signal (image) applies the median technique several times by varying the shape of the window and then taking the median of the median values obtained.
Weak stars and satellites of medium altitude orbit or high altitude orbit satellites are presented as dim point targets in astronomical observation images. They have low intensity and occupy few pixels in the image. In most situations, the sky background is always nonuniform and there are pulse noises. It therefore results in submerged targets in background noise and low signal noise ratio (SNR). The median filter is an effective method to eliminate pulse noise. But you can blur the image after modifying the gray of the no-noise pixels when the image has low SNR. A hybrid medium filter algorithm is proposed to process the SNR image of low weak points. The proposed new method can suppress pulse noises, it can also increase target SNR and does not add additional false targets. The temporal complexity of the median filter is O (N2) (N is the widow's size) and is difficult to implement in hardware. For the application of quasi real time, we propose a rapid implementation of our hybrid medium filter to reduce the complexity of the time to O (N).