Detection of channels in seismic images using the steerable pyramid

Channels have always been important geologic features in the exploration for oil and gas. With 3-D seismic data they can often be mapped easily on time or depth slices. In other situations they can be difficult to detect, due to structural complexity or other factors. There are a number of image-processing algorithms that can be used to enhance linear features such as channels in 3-D seismic volumes. One way involves the use of steerable pyramid filters to partition a seismic image in terms of scale and orientation. Features can then be characterized according to dimensionality and direction using the partitioned image. Here, we explain our implementation of the steerable pyramid in 2-D and 3-D, and show how it can be used to enhance image features. Examples of channel enhancement on synthetic seismic images demonstrate the efficacy of this processing.