A.V. Totsky, I.V. Kurbatov, G.I. Khlopov1, S.I. Khomenko1, V.Ye. Morozov1, J.T. Astola2, and K.O. Egiazarian2
National Aerospace University,
Chkalova Str. 17, 61070, Kharkiv, Ukraine
1A. Usikov Institute of Radio Physics and Electronics,
National Academy of Sciences of Ukraine
12, Academician Proskura Str., Kharkiv 61085, Ukraine
2Tampere University of Technology, Signal Processing Laboratory,
P. O. Box 553, FIN-33101, Tampere, Finland
Extraction of Instantaneous Frequencies from Time-Varying Bispectrum Estimates of Coherent Radar Echo Responses for Moving Objects

Coherent, homodyne, and continuous-wave radar operating with wavelengths of 3 cm, 2cm and 8 mm has been employed for collecting the radar returns obtained from moving objects. Non-stationary, nonlinear frequency modulated and multi-component radar echo responses were analyzed and described as the sum of the Doppler frequency shifted signals (polynomial chirp-like components). Instantaneous frequencies corresponding to the radar radiation scattered from different parts of a moving extended target were extracted from the short time-varying sequences of bimagnitude estimates and mapped into the time-frequency domain. Experimental investigations of the approach demonstrate clean recovery of evolutionary phase coupled harmonics for the test targets a swinging metallic sphere, and a walking person. The obtained time-frequency distributions can be used in automatic radar target recognition systems to reveal more information with new features.

Back
L.J. Morales-Mendoza1, R.F. Vazquez-Bautista1, J.A .Andrade-Lucio2, and O.G. Ibarra-Manzano2
1CINVESTAV-IPN, Prolong. Lopez Mateos Sur No. 590, Gdl., Jalisco, Mexico.
2Electronics Dep., Guanajuato Univ., 912, C. P. 36730, Salamanca Gto, Mexico
Regularization and Enhanced in Radar Images Via Fusing the Maximum Entropy and Variational Analysis Methods (MEVA)

In this article, we present a new fusion strategy for aggregating both the regularization and the anisotropic diffusion paradigms in radar mages reconstruction. The fusion is mainly addressed to gain the highlight features that are involved, in this case, the robust error norm for Variational Analysis (VA) method and the regularized Maximum Entropy (ME) method-based degrees of freedom. The fused method is so-called the Maximum Entropy–Variational Analysis method (MEVA). The method is developed and computational implemented using the modified Hopfield neural network. Furthermore, we present several selected computer simulation examples where real images are addressed to illustrate the outstanding usefulness of this method.

Back
Gerardo Trejo-Caballero, Victor Ayala-Ramirez, Arturo Perez-Garcia, and Raul E. Sanchez-Yanez
Universidad de Guanajuato FIMEE, Tampico No. 912, 36730 Salamanca, MEXICO
Edge Detection in Real Images Using a Neural Network Approach

We propose a neural network approach used in real images to detect edges. Our net detects edges in different directions by analyzing the neighborhood pixels for each pixel of the image. In this work, we show the results of our approach as well as a qualitative comparison with other available algorithms in the literature.

Back
Jose Ruiz-Pinales, Juan Jorge Acosta-Reyes, and Rene Jaime-Rivas
University of Guanajuato/FIMEE, 36730 Salamanca, Gto. Mexico
Feature Extraction for Support Vector Machines Face Recognition

In this paper, we present a feature set selection methodology for the recognition of upright faces by means of support vector machines. We perform face recognition in 3 stages: feature extraction, sub-sampling and classification. We employ a measure based on mutual information in order to select the feature sets. The probabilities required for the computation of the mutual information are estimated by using support vector machines.

Back
Mario A. Ibarra-Manzano, J. Gabriel Avina-Cervantes, Dora L. Almanza-Ojeda, and Jose Ruiz-Pinales
University of Guanajuato. Facultad de Ingenieria Mecanica,
Electrica y Electronica.Av. Tampico No. 912. Salamanca, Guanajuato, 36730, Mexico.
Detection of Closure Features in Gray-Level Images by Using Support Vector Machines

This paper describes an optimal architecture to detect closure features in gray-level images by using Support Vector Machines. This scheme has been extensively tested to recognize some handwriting discriminant features as loops, closure, etc. Closure feature extraction is represented by spatial relation between each tested point and its neighbors. Database for all closure and non-closure features is built. The database allows us to train the SVM with a Gaussian kernel. The proposed method is a fast way to detect handwriting features and it has been extensively tested in the context of closure detected with very reliable results.

Back
Jose Ruiz-Pinales1, Rene Jaime-Rivas1, and Maria Jose Castro2
1FIMEE – University of Guanajuato, Tampico 912, 36730 Salamanca, Gto., Mexico.
2Departamento de Sistemas Informaticos y Computacion - Universidad Politecnica de Valencia,
Discriminative Capacity of Perceptual Features in Handwriting Recognition

This work presents an analysis of the perceptual features, those containing most of the discriminative information of a cursive word in off-line handwriting recognition. A extraction method based on Hough Transform, but preserving local spatial information is presented. The method has been proved with a holistic system, which works on gray-level images and convolutional neural networks.

Back
A.V. Gritsunov
Kharkiv National University of Radio Engineering and Electronics,
14, Lenin Ave, Kharkiv, 61166, Ukraine
On the Reasons for Noises in Cross-Field Devices

The causes of noises in both static and the dynamic modes of the cross-field devices (CFD) are investigated using the method of computing experiments. In the static mode they are: the tangential eigenmode oscillations of the electron cloud; the secondary emission bunching of the core; the solitons in the electron flux. In the dynamic mode they are the transient and the stable convective clouds in the electron spokes. An influence of the abovementioned instabilities on the electron flux oscillation spectra in CFD is considered.

Back
V.G. Sugak, A.V. Bukin, O.A. Ovchinkin, Yu.A. Pedenko, and Yu.S. Silaev
A. Usikov Institute of Radio Physics and Electronics,
National Academy of Sciences of Ukraine
12, Academician Proskura Str., Kharkiv 61085, Ukraine
Use of the Georadar for Groundwater Table Determination and Mappings of the Territories Polluted with Oil

The results of theoretical and experimental research of the possibilities of subsurface radar probing with the aim of solving the problems of engineering and geological examinations – determination of the groundwater depth; defining the lithologic composition of the soils in the zone of aeration as well as the depth of occurrence and the thickness of oil product blanket formed as a result of various sort of man-caused leakages are provided. It is demonstrated that use of frequency-scanning radar is capable of successful solving of the above problems providing for a number of essential advantages compared to both traditional methods of geophysical researches and to the use of subsurface probing radars with pulse signals.

Back
M.V. Lyashenko
Institute of Ionosphere MES and NAS of Ukraine
16, Krasnoznamyonnaya Str., 61002 Kharkiv, Ukraine
Simulation of Seasonal Variations of Electron Concentration in the Ionospheric F2-Peak for Midnight and Noon Local Time

The experimental database obtained at Kharkiv incoherent scattering radar for the period from 1986 till 2002 had been used for modeling seasonal variations of the electron concentration Nm in the ionospheric F2-peak. The regression dependencies of the Nm upon the solar activity index were built. Model coefficients of regression are obtained for every month. Seasonal variations of Nm for midnight and noon local time and for different levels of solar activity were recovered. A comparison between calculated and experimental values was carried out. Good matching between them at any solar activity level was achieved. The particularities of seasonal variations for various phases of solar activity are considered. The obtained results are used for upgrading the empirical ionospheric model of the Central European region.

Back