Cone-shaped volcanoes have important research significance and application value due to their typical cone shape and unique structural features. The existing methods for recognizing volcanoes are mainly morphological feature matching and machine learning. In general, the former has low recognition accuracy, while the latter requires a large number of training samples. The contour lines of cone-shaped volcanoes are distributed in concentric circles. Furthermore, from the center outwards, the elevation of the contour lines increases first and then decreases. Based on the morphological characteristics of cone-shaped volcanoes and the Hough transform algorithm, the main algorithm includes (1) preliminary filtering of contour lines, (2) filtering circular contour lines based on random Hough transform, (3) grouping contour lines based on contour trees, (4) recognizing cone-shaped volcanoes based on concentric-circle contour lines, and (5) automatically mapping cone-shaped volcanoes. Case studies demonstrate the effectiveness of this method for detecting cone-shaped volcanoes in the Western Galapagos shield volcanoes and the Mariana Trench submarine volcano group. The proposed algorithm has low missed and false alarm rates, which is basically consistent with the manual recognition results. This method can effectively automatically recognize cone-shaped volcanoes and cone-shaped landscapes and is a powerful means to support deep-space and deep-sea exploration.