Shallow channels in the South Caspian Sea are major hazards which need to be identified on seismic data prior to well location positioning and drilling operation. Dip-steered-median filtering and automatic gain control were applied to enhance the quality of seismic data during pre-processing. Edge-sensitive attributes such as similarity and curvature were calculated on a migrated poststack channelized seismic volume from the South Caspian Sea. Afterwards, the acquisition footprints along the inline direction which interfered with channel edge detection on seismic attributes were de-striped based on factorial kriging. Moreover, shearlet edge detection algorithm as a novel edge-sensitive seismic attribute were introduced and implemented. Besides, a dense set of horizons (HorizonCube) were extracted after manually picking three horizons in the cube. The existence of geohazards through geo-time slicing based on a flattened seismic cube was also investigated. Furthermore, red-green-blue (RGB) blending of spectral decomposition components to visualize thickness variation within the channels was performed. In the last step, the bodies of channels were successfully extracted using Thalweg tracker.