An automated lineament detection method based on a weighted Hough transform is presented. The method utilizes the result of unsupervised classification based on Kohonen’s self-organizing maps, for vector quantizing the input data space, followed by neuron clustering. It then post-processes the classification result with classical image processing techniques and finally applies the modified Hough transform in order to identify lineaments. The capabilities of the method are described using geophysical (airborne magnetic and electromagnetic) data from the Vammala area in Finland. The results of the automated analysis show major geological faults in the selected area. Finally, comparisons with the classical Hough transform algorithm show the advantages of our proposed modifications.