Application of the Hough transform for automated interpretation of linear features in imageries

The use of satellite images for evaluation of structural geology is traditionally undertaken by geologists who manually work through lineament maps as a basis for their further interpretations. The application of the Hough transform as an automated detection of linear features has been discussed by several authors. They have demonstrated that this algorithm has the potential of detecting lines and linear features on images, and thereby allows automated mapping of lineaments for further interpretation of structural geology. The present paper describes an application of the Hough transform for the interpretation of digitally enhanced satellite imageries covering an area of central Norway. The processing of the images includes directional filtering and a maximum filter prior to segmentation of structural features. A novel algorithm based on the Hough transform combined with a moving window algorithm has been implemented and applied for the classification of lineaments according to directionality and length. Regional and local variations of linear features emerge by using different window sizes and "step lengths". The results of the automatic analysis are compared with published interpretations of the structural geology of the investigated area, and it is concluded that the relevant geological features can be detected by this automatized method.