| Abstract: | A novel approach of digital image processing technology is applied to improve SNR of seismic images. At first,we analyze the characters of line-like texture in seismic images, and then a preprocessing method named 2 D tracing horizon filtering is designed. After that, the technology of optical flow analysis is adopted to calculate the displacement vectors of adjacent pixels between neighboring seismic images. At last, the novel image accumulation algorithms are proposed, which are applied to greatly improve SNR and definition of seismic images. The experimental results show that SNR of seismic section images with SNR of about 20 dB and 17 dB are increased 8 dB～9 dB under keeping signal energy 67%～80% by processing section images and 3dB～4dB under keeping signalenergy 80～90% by processing horizontal slice images. Thereby, the proposed novel approaches are very helpful to the correct seismic interpretation and have very important significance for seismic exploring. |
