Système de tri couleur par capteur flou
Color sorting system by fuzzy sensor
OPEN ACCESS
This article presents a color recognition system formalized under the fuzzy sensor concept. The main objective is to improve the color classification of wooden boards. Our study relates the development of an industrial vision system allowing the recognition of gradual colors. This context imposes a strong reliability constraint, because the currently used sensors are not always enough robust. Then, we are interested in techniques which improve the taking into account of the physical measure imprecision and the uncertainty concerning the definition of the wood color by a Human, the color classes being always neither identified nor separated. Moreover, the different users can have different perceptions of them. Besides, the vision system must be easy to tune. To carry out such a system, we propose to base the fuzzy sensor on a classification method with fuzzy linguistic rules (Fuzzy Reasoning Classifier) which main advantages reside in its generalization capacity from small training data sets and in the interpretability of its rule set. The obtained results show the efficiency of our intelligent sensor.
Résumé
Cet article présente un système de reconnaissance couleur formalisé sous le concept de capteur flou. L’objectif principal est d’améliorer la classification couleur de planches de bois. Notre étude concerne le développement d’un système de vision industriel permettant la reconnaissance de couleurs graduelles. Ce contexte impose une contrainte forte de fiabilité, les capteurs utilisés aujourd’hui n’étant pas toujours suffisamment robustes. Ainsi, nous nous sommes intéressés à des techniques qui améliorent la prise en compte des imprécisions des mesures physiques et la subjectivité concernant la définition de la couleur du bois par l’Homme, les classes de couleur n’étant jamais bien identifiées ni séparées. De plus, les différents utilisateurs peuvent en avoir des perceptions différentes. Par ailleurs, et dans ce contexte particulier, un système de vision doit être simple à régler. Pour réaliser un tel système, nous proposons de baser le capteur flou sur une méthode de classification par règles linguistiques floues (Fuzzy Reasoning Classifier) dont les principaux avantages résident dans sa capacité de généralisation à partir de lot de données réduits en apprentissage et dans l’interprétabilité de sa base de règles. Les résultats que nous obtenons montrent l’efficacité de notre capteur intelligent.
Fuzzy sensor, Color measurement, Pattern recognition, Fuzzy rules, Image processing
Mots clés
Capteur flou, Mesure couleur, Reconnaissance de formes, Règles floues, Traitement d’images
[Albus] J.S. ALBUS, «Outline for a theory of intelligence», IEEE trans. On SMC, Vol. 21, 1991, p. 473-509.
[Alcala] R. ALCALA, J. ALCALA-FDEZ, F. HERRERA, J. OTERO, «Genetic learning of accurate and compact fuzzy rule based systems based on the 2-tuples linguistic representation », Int. Journal Of Approximate reasoning, Vol. 44, 2007, p. 45-64.
[Benoit 1] E. BENOIT, L. FOULLOY, «High functionalities for intelligent sensors, application to fuzzy colour sensor », Measurement, Vol. 30, 2001, p. 161-170.
[Benoit 2] E. BENOIT, L. FOULLOY, «Symbolic sensors: one solution to the numerical-symbolic interface», Proc. Of the IMACS DSS&QR workshop, Toulouse, France, 1991, p. 321-326.
[Berthold] M.R. BERTHOLD, «Mixed fuzzy rule formation», Int. Journal. Of Fuzzy Sets and Systems, Vol. 32, 2003, p. 67-84.
[Blake] C. BLAKE, E. KEOGH, C.J. MERZ, «UCI Repository of machine learning databases», University of California, Irvine.
[Bombardier] V. BOMBARDIER, C. MAZAUD, P. LHOSTE, R. VOGRIG, «Contribution of Fuzzy Reasoning Method to knowledge Integration in a wood defect Recognition System », Computers in Industry Journal, vol 58, p. 355–366, 2007.
[Bouchon-Meunier] B. BOUCHON-MEUNIER, La logique floue et ses applications, Ed. Addison-Wesley, 1995.
[Burd] N.C. BURD, A.P. DOREY, «Intelligent transducers», Journal of Microcomputer Applications, Vol. 7, 1984, p. 87-97.
[Ciame-Afcet] CIAME-AFCET, «Livre blanc : Les capteurs intelligents – pensée de l’utilisateur», 1987, 169 pages.
[CIE] International Commission on Illumination, Colorimetry, 2nd Edition, Publication CIE No 15.2, 1986.
[Cordon] O. CORDON, M.J. DEL JESUS, F. HERRERA, «A proposal on reasoning methods in fuzzy rule-based classification systems », Int. Journal Of Approximate reasoning, Vol. 20, 1999, p. 21-45.
[Dubois 1] D. DUBOIS, H. PRADE, «On the use of aggregation operations in information fusion processes », Fuzzy Sets and Systems, Vol. 142, 2004, p. 143-161.
[Dubois 2] D. DUBOIS, H. PRADE, «What are Fuzzy rules and how to use them », Fuzzy Sets and Systems, Vol. 84, 1996, p. 169-185.
[Dubois 3] D. DUBOIS, et H. PRADE, « The three semantics of fuzzy sets », Fuzzy Sets and Systems, Vol. 90, p. 141-150, 1997.
[Dubois 4] D. DUBOIS and H. PRADE, «Fuzzy rules in knowledge-based systems – Modelling gradedness, uncertainty and preference », An introduction to fuzzy logic application in intelligent systems, p. 45-68, Kluwer, Dordrecht, 1992.
[Dubois 5] D. DUBOIS, H. PRADE and L. UGHETTO, «Checking the coherence and redundancy of fuzzy knowledge bases », IEEE Trans. Fuzzy Systems, vol. 5, p. 398-417, 1997.
[Fagin] R. FAGIN, « Combining Fuzzy Information from Multiple Systems», Jour. of Computer and System Sciences, Vol. 57, 1999, p. 83-99.
[Hanbury] A. HANBURY, «Morphologie mathématique sur le cercle unité avec applications aux teintes et aux textures orientées», Thèse de l’École Nationale Supérieure des Mines, Paris, 2002.
[Hao] P.Y. HAO, J.H. CHIANG, Y.K. TU, «Hierarchically SVM classification based on support vector clustering method and its application to document categorization», Expert Systems with Applications, Vol. 33, 2007, p. 627-635.
[Hudelot] C. HUDELOT, J. ATIF, et I. BLOCH, « Ontologie de relations spatiales floues pour l’interprétation d’images. », Rencontres Francophones sur la Logique Floue et ses Applications – LFA 2006, Toulouse, France, pp.363-370, 2006.
[Ishibuchi 1] H. ISHIBUCHI, K. NOZAKI, H. TANAKA, «Construction of fuzzy classification systems with rectangular fuzzy rules using genetic algorithms», Fuzzy Sets and Systems, Vol. 65, 1994, p. 237-253.
[Ishibuchi 2] H. ISHIBUCHI, K. NOZAKI, H. TANAKA, «Distributed representation of fuzzy rules and its application to pattern classification », Fuzzy Sets and Systems, Vol. 52, 1992, p. 21-32
[Ishibuchi 3] H. ISHIBUCHI AND T. NAKASHIMA, «Performance evaluation of fuzzy classifier systems for multi-dimensional pattern classification problems», IEEE Trans. Systems, Man and Cybernetics – Part B: Cybernetics, vol.29, p. 601–618, 1999.
[Kline] D.E KLINE, R.W. CONNERS, P.A. ARAMAN, «Technology to Sort Lumber by Color and Grain for Furniture Parts », Conference: Quality Lumber Drying in the Pacific Northwest, p. 67-73, 1999.
[Lee] P.S. LEE, A.L. DEXTER, «A fuzzy sensor for measuring the mixed air temperature in airhandling units », Measurement, Vol. 37, 2005, p. 83-93.
[Leon] K. LEON, D. MERY, F. PEDRESCHI, J. LEON, «Color measurement in L*a*b* units from RGB digital images », Food research international, Vol. 39, 2006, p. 1084-1091.
[Lu] Q. LU, «A Real-Time System for Color Sorting Edge-Glued Panel Parts », Thesis of the Faculty of the Virginia Polytechnic Institute and State University, Blacksburg, Virginia, December 1997.
[Maloney] L.T MALONEY, B.A. WANDELL, «Color constancy: a method for recovering surface spectral reflectance», Journal of the Optical Society of America A, Vol. 3, No. 1, 1986, p. 29-33.
[Malamas] E.N. MALAMAS, E.G.M. PETRAKIS, M. ZERVAKIS, L. PETIT, J-D. LEGAT, «A survey of industrial vision systems, applications and tools », Image and Vision Computing, Vol. 21, 2003, p. 171-188.
[Marsala] C. MARSALA, «Fuzzy decision trees to help flexible querying », Kybernetika, vol. 36, 2000, p. 689-705,
[Marszalec] E. MARSZALEC, M. PIETIKAINEN, « Some aspects of RGB vision and its applications to industry. », International Journal of Pattern Recognition and Artificial Intelligence, Vol. 10, 1996, p. 55-72.
[Mauris 1] G. MAURIS, V. LASSERRE, L. FOULLOY, «A fuzzy approach for the expression of uncertainty in measurement», Measurement, Vol. 29, 2001, p. 165-177.
[Mauris 2] G. MAURIS, E. BENOIT, L. FOULLOY, «Fuzzy Linguistic Methods for the Aggregation of Complementary Sensor Information», Aggregation and Fusion of Imperfect Information, 1998, p. 215-230.
[Mauris 3] G. MAURIS, E. BENOIT, L. FOULLOY, «The aggregation of complementary information via fuzzy sensors », Measurement, Vol. 17, 1996, p. 235-249.
[Mendel] J.M. MENDEL, «Fuzzy logic systems for engineering: A tutorial», Proceedings of the IEEE, vol. 83, no. 3, p. 345–377, 1995.
[Michie] D. MICHIE, D.J. SPIEGELHALTER, C.C. TAYLOR, «Machine Learning Neural and Statistical Classification», Ellis Horwood, 1994.
[Nakoula] Y. NAKOULA, S. GALICHET, et L. FOULLOY, « Learning of a fuzzy symbolic rule base », Proc. Of 3rd European Congress on Intelligent Techniques and Soft Computing, Aachen, Allemagne, 1995.
[Nozaki] K. NOZAKI, H. ISHIBUCHI, H. TANAKA, «A Simple but powerful heuristic method for generating fuzzy rules from numerical data », Fuzzy Sets and Systems, Vol. 86, 1997, p. 251-270.
[Perez Oramas] O. PEREZ ORAMAS, «Contribution à une méthodologie d’intégration de connaissances pour le Traitement d’images. Application à la détection de contours par règles linguistiques », Thèse de l’Université Henri Poincaré, CRAN, CNRS UMR 7039, Nancy, 2000.
[Pham] D.T. PHAM, S. SAGIROGLU, «Training multilayered perceptrons for pattern recognition: a comparative study of four training algorithms», International Journal of Machine Tools & Manufacture, Vol. 41, 2001, p. 419-430.
[Philipp] I. PHILIPP, T. RATH, «Improving plant discrimination in image processing by use of different colour space transformations », Computers and electronics in agriculture, Vol. 35, 2002, p. 1-15.
[Sangwine] S.J. SANGWINE, R.E.N. HORNE, The colour image Handbook, Ed. Chapman and Hall, 1998.
[Smeulders] A.W.M. SMEULDERS, M. WORRING, S. SANTINI, A. GUPTA, and R. JAIN, «Contentbased image retrieval at the end of the early years», IEEE Trans. PAMI, vol. 22, p. 1349-1380, 2000.
[Schmitt 1] E. SCHMITT, C. MAZAUD, V. BOMBARDIER, P. LHOSTE, « A Fuzzy Reasoning Classification Method for Pattern Recognition », Proc of the 15th International Conference onFuzzy Systems (FUZZIEEE’06), Vancouver, Canada, 2006, p. 5998-6005.
[Schmitt 2] E. SCHMITT, V. BOMBARDIER, P. CHARPENTIER, «SelfFuzzification Method according to Typicality Correlation for Classification on tiny Data Sets», Proc. 16th Int. Conf. on Fuzzy Systems, FUZZIEEE'07, Londres, Angleterre, 2007, p. 1072-1077.
[Schmitt 3] E. SCHMITT : « Contribution au Système d’Information d’un Produit Bois. Appariement automatique de pièces de bois selon des critères de couleur et de texture », Thèse de l’Université Henri Poincaré, CRAN, CNRS UMR 7039, Nancy, 2007
[Schmitt 4] E. SCHMITT, V. BOMBARDIER, L. WENDLING, «Improving Fuzzy Rule Classifier by Extracting Suitable Features from Capacities with Respect to the Choquet Integral», IEEE trans. On System, man and cybernetics- part B, Vol 38, N° 5, October 2008.
[Srikanteswara] S. SRIKANTESWATRA, Q. LU, W. KING, T. DRAYER, R. CONNERS, E. KLINE, P. ARAMAN, «Real-time implementation of a color sorting system », SPIE, Vol. 3205, 1997.
[Sugeno] M. SUGENO, «An introductory survey of fuzzy control », Information Sciences, vol. 36, p. 59-83, 1985.
[Zadeh 1] L.A. ZADEH, «The concept of linguistic variable and its application to approximate reasoning », Information sciences, Vol. 8, 1975, p.199-249.
[Zadeh 2] L.A. ZADEH, «Outline of a new approach to the analysis of complex systems and decision processes », IEEE trans. On Systems, Man and Cybernetics, Vol. SMC3, 1973, p. 28-44.
[Zadeh 3] L.A. ZADEH, «Fuzzy sets », Information and control, Vol. 8, 1965, p. 338-353.