Applied Soft Computing 13 (2013) 4738–4749 Contents lists available at ScienceDirect Applied Soft Computing j ourna l ho me page: www.elsevier.com/locate /asoc Linguistic description about circular structures of the Mars’ surface Daniel Sanchez-Valdes ∗ , Alberto Alvarez-Alvarez, Gracian Trivino European Centre for Soft Computing, C/ Gonzalo Gutierrez Quirós, s/n, 33600 Mieres, Asturias, Spain a r t i c l e i n f o Article history: Received 2 August 2012 Received in revised form 10 June 2013 Accepted 5 August 2013 Available online 3 September 2013 Keywords: Linguistic description of data Computing with perceptions Image description Granular linguistic model of phenomena a b s t r a c t Satellites situated in the orbit of Mars have provided and continue providing thousand of images of the planet surface. Nevertheless the number of expert geologists analyzing these images is limited. Typically, these experts provide linguistic descriptions of their observations remarking the relevant features in the image and ignoring the irrelevant details for a given goal. In this paper, we apply our research in the field of Computational Theory of Perceptions to the challenge of developing computational systems able to generate linguistic reports comparable with the ones provided by human experts. We present a description of our contribution to solve this problem including last results of our research in this field. For example, we explore how to represent the multidimensional domain of computational perception values. We develop up the use of the relevance as an attribute of perceptions that allows us to generate reports that are automatically suited according to the user goals. We provide an application example as a demonstration of concept. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Current technologies allow us to acquire and store big amounts of data about complex phenomena in many areas of science and technology. In order to be useful, these data must be interpreted and represented in an understandable way, by giving in each type of situation their relationship to data context and, in general, the information related to each specific phenomenon. Currently, these types of descriptions are reports that contain text and graphics pro- duced by human experts. Nevertheless, the relation between the amount of data to analyze and the number of experts available is growing dramatically. This situation causes a strong demand for computational systems that can interpret and describe linguisti- cally the large amount of information that is being generated in many areas. This paper was motivated by the existence of a huge database of thousands of images of the Mars’ surface produced by satellites and it is part of a collaboration project with the Spanish National Institute for Aerospace Technology (INTA). These images, which are usually analyzed by a small number of available experts on Martian geology, contain important information. For example, one of the most important research topics about the Mars planet is the seek of water, which can be developed by the analysis of its surface. In [1], authors use infrared to visible wavelength images of the Mars’ ∗ Corresponding author. Tel.: +34 985456545. E-mail addresses: daniel.sanchezv@softcomputing.es (D. Sanchez-Valdes), alberto.alvarez@softcomputing.es (A. Alvarez-Alvarez), gracian.trivino@softcomputing.es (G. Trivino). surface to obtain relevant geological information that may help in the location of water frost. In [2], authors analyze the measure- ments of elevations in order to extract a highly accurate global map of the topography of Mars that determines an upper limit of the present surface water inventory. In this work, we will use digital image processing techniques [3,4], in order to obtain automatically information about the pres- ence of circular structures of the Mars’ surface and produce a linguistic description of it. There are several examples related to this work. In [5], authors use 3D image processing to automate the surface quality analysis of exterior car body panels. From the viewpoint of describing images in natural language (NL), there are recent works based on the use of Fuzzy Logic (FL) such as [6], where authors propose a hierarchical fuzzy segmentation of the image and a collection of linguistic features able to describe each region. In [7], authors explain how a fuzzy object-relational database man- agement system can be employed to implement and integrate the different elements needed for the linguistic description of images, namely, ontology, concept representation and language generation. Our research is based on the Computational Theory of Per- ceptions (CTP) introduced by Zadeh [8–10]. In previous works in this line, we have developed computational systems able to gen- erate linguistic descriptions of different types of phenomena. For example, we generated financial reports from data taken from the Spanish Securities Market Commission (CNMV) [11]. We have also worked in the field of activity recognition [12] and gait analysis [13], by providing automatically generated linguistic reports based on accelerometer data. And, we have also dealt with data related to the field of Intelligent Transportation Systems (ITS), where we gen- erated assessing reports in truck driving simulators [14,15]. In [16], 1568-4946/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.asoc.2013.08.003