Anna Rampini

A fuzzy set-based accuracy assessment of soft classification

Despite the sizable achievements obtained, the use of soft classifiers is still limited by the lack of well-assessed and adequate methods for evaluating the accuracy of their outputs. This paper proposes a new method that uses the fuzzy set theory to extend the applicability of the traditional error matrix method to the evaluation of soft classifiers. It is designed to cope with those situations in which classification and/or reference data are expressed in multimembership form and the grades of membership represent diAerent levels of approximation to intrinsically vague classes. ” 1999 Elsevier Science B.V. All rights reserved.

Slope Instability Zonation: a Comparison Between Certainty Factor and Fuzzy Dempster–Shafer Approaches

This paper presents a comparison between two methodologies for the evaluation of slope instability and the production of instability maps, using a probabilistic approach and a hybrid possibilistic and credibilistic approach. The first is the Certainty Factor method, and the second is based on Fuzzy Logic integrated with the Dempster–Shafer theory. These methodologies are applied to the 1 : 50,000 scale Fabriano (Marche, Italy) geological map sheet. The results are represented as histograms where the accuracy of the prediction is shown, and the comparison of the results of the methods is discussed.

Image registration by recognition of corresponding structures

A method for automatic image registration which is characterized by its insensitivity to scaling, rotation, and intensity changes is described. The method is based on similarity assessment of the structures in the images and on a check of their spatial arrangement. Pairs of structures that correspond to each other provide sets of control points to geometric mapping functions. An application of the method to remote-sensing image alignment with a reference map is presented. >

Comparison of the multilayer perceptron with neuro-fuzzy techniques in the estimation of cover class mixture in remotely sensed data

Mixed pixels are a major source of inconvenience in the classification of remotely sensed data. This paper compares MLP with so-called neuro-fuzzy algorithms in the estimation of pixel component cover classes. Two neuro-fuzzy networks are selected from the literature as representatives of soft classifiers featuring different combinations of fuzzy set-theoretic principles with neural network learning mechanisms. These networks are: 1) the fuzzy multilayer perceptron (FMLP) and 2) a two-stage hybrid (TSH) learning neural network whose unsupervised first stage consists of the fully self-organizing simplified adaptive resonance theory (FOSART) clustering model, FMLP, TSH, and MLP are compared on CLASSITEST, a standard set of synthetic images where per-pixel proportions of cover class mixtures are known a priori. Results are assessed by means of evaluation tools specifically developed for the comparison of soft classifiers. Experimental results show that classification accuracies of FMLP and TSH are comparable, whereas TSH is faster to train than FMLP. On the other hand, FMLP and TSW outperform MLP when little prior knowledge is available for training the network, i.e., when no fuzzy training sites, describing intermediate label assignments, are available.

A linguistic decision making approach to assess the quality of volunteer geographic information for citizen science

The paper analyses the challenges and problems posed by the use of Volunteered Geographic Information (VGI) in citizen science and a proposal is formulated for assessing VGI quality based on a linguistic decision making approach so as to allow its feasible use for scientific purposes. VGI quality is represented by indicators at distinct levels of granularity which take into account the distinct components of the VGI items. The quality indicators represent both the extrinsic quality, depending on the characteristics and reputation of the sources of information; the intrinsic quality, depending on the distinct accuracy and precision of information; and, last but not least, the pragmatic quality, depending on the user needs and intended purposes. In order to assess the pragmatic quality of VGI items, a linguistic decision making approach is defined that allows users to rank and finally filter the VGI items based on the satisfaction of distinct criteria expressed by means of both linguistic terms, defining soft constraints on the distinct quality indicators, and linguistic aggregators, defining fuzzy operators which combine the satisfaction degrees of the soft constraints at distinct hierarchical levels to yield the final satisfaction of the VGI items. Finally, an example of quality assessment in a glaciological citizen science project is discussed.

Fuzzy reasoning approach to similarity evaluation in image analysis

In image analysis, the concept of similarity has been widely explored and various measures of similarity, or of distance, have been proposed that yield a quantitative evaluation. There are cases, however, in which the evaluation of similarity should reproduce the judgment of a human observer based mainly on qualitative and, possibly, subjective appraisal of perceptual features. This process is best modeled as a cognitive process based on knowledge structures and inference strategies, able to incorporate the human reasoning mechanisms and to handle their inherent uncertainties. This articlea proposes a general strategy for similarity evaluation in image analysis considered as a cognitive process. A salient aspect is the use of fuzzy logic propositions to represent knowledge structures, and fuzzy reasoning to model inference mechanisms. Specific similarity evaluation procedures are presented that demonstrate how the same general strategy can be applied to different image analysis problems.

Fuzzy decision making in the classification of multisource remote sensing data

Our objective was to develop a knowledge-based strategy for the classification, considered a cognitive process, of multisource data including remote sensing images. The main feature of our approach is the use of fuzzy sets as the representation framework. This strategy supports two supervised image classification procedures, one based on a fuzzy statistical classifier and the other on a feed-forward fuzzy trained neural network. Approximate reasoning techniques, based on fuzzy production rules, are applied to model the multifactorial evaluation process in which results from the classification of remote-sensing images are integrated with other data. An example of multisource remote-sensing data classification applied in fire prevention is presented together with numerical results and an experimental verification of the approach.

Development of a satellite remote sensing technique for the study of alpine glaciers

Abstract This paper presents an experiment in MSS image interpretation for the systematic observation of glacier surfaces in the Alps. The glacier monitoring potential of visible and near-infrared data is applied here to the surveying of mountain glaciers, characterized by small areas and strong shading because of their typical morphology. The method developed identifies glacier surfaces by evaluating the intensity values of visible images combined with clearness conditions related to exposure, slope and the surface homogeneity of the glacier. Conditions of clarity, in the absence of a digital terrain model, are estimated from the number of saturated pixels in the visible bands. At a higher level, the near-infrared data used to identify snow and ice surfaces inside the glacier boundaries. The paper discusses the performance of the technique developed, as applied to the analysis of a temporal series concerning a group of 11 small glaciers with critical exposure conditions. The results are expressed as area...

Glacial retreat in the 1980s in the Breonie, Aurine and Pusteresi groups (eastern Alps, Italy) in Landsat TM images

Abstract In the Italian Alps, the majority of glaciers are small (less than 1 km2), but cover a significant glaciated surface area, despite the fact that ground surveys have only dealt with the major glaciers. Analysis of some Landsat TM images taken at the end of the ablation season highlighted retreat modes in the 1980s on all glaciers in the Breonie, Aurine and Pusteresi groups (eastern Alps)—an area in which, in the early 1980s, large glaciers were advancing and small ones receding. A fuzzy set procedure used to identify glacier surfaces was based on integrated use of Landsat TM images and topographic data in a multisource decision making scheme (applied to bands 1, 3 and 5) to combine information derived from elevation, exposure and morphological aspects. Between 1985 and 1987, the loss of glaciated surface was almost 4 km2 (total area 33.04 km2), the greatest retreat being found for the largest glaciers (>1 km2). Between 1987 and 1989, the loss was less than 1 km2 and mostly involved small glaciers ...

Downstream Services for Rice Crop Monitoring in Europe: From Regional to Local Scale

The ERMES agromonitoring system for rice cultivations integrates EO data at different resolutions, crop models, and user-provided in situ data in a unified system, which drives two operational downstream services for rice monitoring. The first is aimed at providing information concerning the behavior of the current season at regional/rice district scale, while the second is dedicated to provide farmers with field-scale data useful to support more efficient and environmentally friendly crop practices. In this contribution, we describe the main characteristics of the system, in terms of overall architecture, technological solutions adopted, characteristics of the developed products, and functionalities provided to end users. Peculiarities of the system reside in its ability to cope with the needs of different stakeholders within a common platform, and in a tight integration between EO data processing and information retrieval, crop modeling, in situ data collection, and information dissemination. The ERMES system has been operationally tested in three European rice-producing countries (Italy, Spain, and Greece) during growing seasons 2015 and 2016, providing a great amount of near-real-time information concerning rice crops. Highlights of significant results are provided, with particular focus on real-world applications of ERMES products and services. Although developed with focus on European rice cultivations, solutions implemented in the ERMES system can be, and are already being, adapted to other crops and/or areas of the world, thus making it a valuable testing bed for the development of advanced, integrated agricultural monitoring systems.