Research Of NRM Image Noise Reduction Processing

NRM imaging is not damaging to the patient, and also has the very high accuracy for the diagnosis of the disease. But its scan time is much long and imaging speed is low. Compressed sensing theory is far lower than the Nyquist sampling frequency to sampling samples, shorten the sampling time, reduce the storage space and improve the imaging speed.People apply compression sensing theory to magnetic resonance imaging. In fact, it is the design of the measurement matrix, namely the design of the Fourier under-samp ling. This article emphatically introduces the interested region sampling design. The basic idea is to design a kind of sampling methods, and this method can sample interested areas as much as possible, sampling no interested region as little as possible or no sampling.In the process of NMR imaging, imaging will be affected by noise which reduces the quality of imaging. Based on this, in the case of known noise model, this paper puts forward a new method of noise reduction. By constantly adjusting the parameters of the noise model, we can deal with the k space after the sampling data, filter out the noise of the information, then use Fourier inverse transformation for image reconstruction. With gaussian filtering noise reduction on spatial domain and gaussian low-pass filtering noise reduction method on frequency domain, we find that the method get the ideal result.