Millimeter-Wave Weapons Detection System has been widely used in airport,custom-house and other safety checkpoint in recent years.In this paper,we propose a new method to make the image produced by Millimeter-wave Weapons Detection System more readily identifiable.It is found that the hidden weapon in the image is easier to be distinguished after adding suitable noise level and being denoised by 3D transform domain collaborative filtering (BM3D) [1].Furthermore,we proposed a improved version of BM3D to promote the performance.Experiment results are given to demonstrate the different performance for different added noise level.