Ladislav Beran
University of Pardubice
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Ladislav Beran.
international conference radioelektronika | 2015
Pavel Chmelar; Ladislav Beran; Nataliia V. Kudriavtseva
Lasers are widely used in measurement systems, health care and in electrical engineering in general. One recent area in measurement represents 3D range scanning. The accurate laser line detection is crucial for their precision. In this paper we would like to introduce a new method of the color-based laser line detection by using felicitous Gaussian mixture model (GMM). Detection algorithm process a static image captured by camera. There are several advantages of using this sophisticated model. The color segmentation process is able to deal with saturation, different laser intensity, darker objects and ambient daylight. Various color models it is possible to use in modeling process. For laser color detection we are using two basic color models, RGB and HSV. The paper includes a rigorous comparative evaluation of the two aforementioned color models in a segmentation colored laser line. Results of 3D range scanning are compared and recommendations for using color GMM segmentation are mentioned in last part of this paper.
international conference radioelektronika | 2015
Ladislav Beran; Pavel Chmelar; Lubos Rejfek
Precise navigation of mobile robots in unknown areas is common problematic in many research papers. For navigation in outdoor areas Global Navigation Satellite System (GNSS) is typically used. For better position estimation in Europe augmentation satellite system called European Geostationary Navigation Overlay Service (EGNOS) it is used. In US it is similar system called Wide Area Augmentation System (WAAS). These systems are not usable in indoor areas because the weak of signal and the low position accuracy. For indoor areas, it is necessary to use an alternative navigation system - independent on satellite navigation. In our research we are focusing on visual odometry using a monocular camera. Advanced image processing methods on camera images to achieved accurate visual odometry are applied.
international conference radioelektronika | 2015
Lubos Rejfek; Ladislav Beran; Ondrej Fiser
This paper shown the correction of the radar power in dependence on the target distance. This correction is important for right type estimation of the target. Type of the used radar is frequency modulated interrupted continuous wave radar. This radar receives different length of the pulse for the different distances of the targets and different transmitted pulses. Received powers are analyzed by the power spectral density methods. The Fourier transform is used in this paper. Results from this work are important for the radar reflectivity measurement and these results will be implemented to the real radar system.
conference on microwave techniques comite | 2015
Ladislav Beran; Pavel Chmelar; Lubos Rejfek; Jaroslav Chum; Zbysek Mosna
This electronic document is about comparison of the methods and systems for the measurement of the ionosphere used at the geophysical observatory Pruhonice (50N, 14.5E). Three systems for ionospheric monitoring are installed at the station. The first system for the monitoring of the ionosphere is Digisonde, the second system is dual frequency GNSS receiver and the third system is continuous Doppler sounding. Each system has its benefits and weaknesses which are compared in this paper.
international conference radioelektronika | 2016
Natalija Chmelarova; Pavel Chmelar; Ladislav Beran; Lubos Rejfek
Nowadays lasers in the measurement industry are widely used. The accuracy of optical rangefinders, which detect a laser line directly from a taken frame, lies in precise laser line detection. Unfortunately, during the measurement process some laser line reflections can occur, laser colored objects are present on the scene or segmented laser line is too wide due to the laser saturation effect. These mentioned phenomena have influence on the measurement precision. It is possible to suppress these effects partially by checking the laser line width, but these points are missing in final measurement results. This paper describes a detailed analysis of segmented laser line components for elimination these phenomena, and the final distinguishing, which parts of a segmented laser line are used in measurement process. The paper also includes a comparison with a centroid method, which was used for a laser line center estimation.
international symposium elmar | 2014
Pavel Chmelar; Ladislav Beran; Nataliia V. Kudriavtseva
Point cloud represents the set of measured points. Several methods for point cloud projection have already been proposed. These methods process measured points without angle rotation. We, however, believe that rotations of measurement points simplify whole object detection process. We propose to unfold a measured point cloud of indoor 3D space into single plane with preservation of height dimension. For unfolding of measured points we are using a suitable rotation matrix. This simplification allows to use a simpler methods for object detection. Described algorithm has been verified on several examples - indoor rooms.
international symposium elmar | 2015
Pavel Chmelar; Ladislav Beran; Lubos Rejfek; Nataliia V. Kudriavtseva
Nowadays lasers in the measurement industry are widely used. The accuracy of optical rangefinders, which detect a laser line directly from a taken frame, lies in precise laser line detection. By using a lens with a smaller focal length, a barrel distortion can occur. In this paper we would like to introduce an effective method for barrel distortion correction. The algorithm also includes a super-sampling method to prevent line discontinuities especially on image borders, where the influence of distortion effect is significant. The algorithm description also includes substantial speed-up of the whole process by using a correction matrix. A correction matrix is estimated only once for a specific frame dimension and users parameters. Then the camera correction matrix is used in the mapping process. The paper also includes a time comparison between using a different super-sampling quality and the benefit of using the correction matrix.
international symposium elmar | 2017
Lubos Rejfek; Ondrej Fiser; David Matousek; Ladislav Beran; Pavel Chmelar
This electronic document is aimed at the corrections of the received power through frequency modulated interrupted continuous wave radars. These corrections are important for the correct interpretation of the target characteristics. Power corrections, described in this paper, are modified for the measurement of the Doppler shifts. Doppler shifts are calculated by the using of the two dimensional fast Fourier transform. In practical part of this paper four types of windows were compared during calculation of the two dimensional spectra. We concluded that the correction coefficients should be determined for each window individually.
international symposium elmar | 2017
Lubos Rejfek; Ondrej Fiser; David Matousek; Ladislav Beran; Pavel Chmelar
This conference paper is aimed at the analysis of the frequency modulated interrupted continuous wave radar (PCDR35). This radar was developed under the Institute of Atmospheric Physic Czech Academy of Sciences leadership. A primary aim of this radar is the monitoring of meteorological targets (rain cells, hail stones and other). In this paper a radar measurement of the corner reflector with the theoretical reflectivity determination are compared. In the last part of this paper an evaluation of this radar and possible modifications (classical FMCW) are described.
international conference radioelektronika | 2017
Pavel Chmelar; Ladislav Beran; Lubos Rejfek; Natalija Chmelarova
This paper describes the point cloud visualization for rotary optical rangefinders. Measured distance data with image information from captured frames give a possibility to create a colored point cloud. The distance is measured by the vertical laser line and to select a color value directly from an actual measured frame is not possible. For this reason, the previous measurement frame is used. The individual frames shifting is estimated by the temple matching algorithm. The purpose of the paper is to show how easily it is possible to combine measurement data with image information in case of a color point cloud construction. The paper includes several examples of different point clouds to verify the algorithms ability.