Texture Analysis And Its Application In Image Segmentation Of Seismic Sections

The auto-explanation of seismic section images has important utility value to segmentation of seismic texture images. The analysis of texture characters is the most important foundation work of texture segmentation. We get orientation characters of texture from macro-texture and micro-texture then use FCM to cluster. The way of texture segmentation based on Local Binary Pattern(LBP) was brought forward in the research. These method mainly uses LBP to get micro-texture characters and the laws texture energy measure to get macro-texture characters. The computer simulations show it is a good antinoise method. An improving method is proposed because of the complexity of grey value mean value algorithm in laws energy measure.A texture segmentation method based on orientation derivative is proposed. These method use orientation derivative to get the micro-texture characters and the laws of texture energy measure to get six-dimensional eigenvector and it is classified by FCM to finish segmentation. The method is simple and easy to achieve, it has little mistakes on the brim.Different frequency characters of texture images are picked up by designing band-pass filter and Gabor filter on space/frequency and clustered by FCM. The computer simulations show this method has good effect.After analyzing seismic section images deeply, we apply the methods we talked above on segmentation of seismic section images. The methods are proved availably.