Various levels and types of noise are responsible for hindering the valuable information obtained through shallow depth geophysical exploration of archaeological sites. Wavelet transform techniques were tested as a method for decomposition of the original geophysical data in order to eliminate the noise levels inherent to the geophysical measurements. Unsupervised classification techniques were employed for the final fusion of different datasets originating from various surveys or processing procedures. The resulting images were able to enhance the subsurface targets, eliminating the noise levels and exploiting fully the properties of the geophysical techniques used.
