• Title/Summary/Keyword: Frequncy matching

Search Result 3, Processing Time 0.015 seconds

FREQUENCY SELECTIVE RECURSIVE LP OF HARMONIC SPECTRA

  • SeungHyonNam
    • Journal of the Korean Geophysical Society
    • /
    • v.4 no.4
    • /
    • pp.231-238
    • /
    • 2001
  • In this paper, an efficient LP method ofr discrete harmonic spectra is proposed and discussed. A new efficient LP method is a combination of recursive and frequency selective LP. While the recursive LP provides better spectral matching in spectral hills, frequency selective LP eliminates numerical instability and improves spectral matching when the harmonics are confined in the low frequncy region. The proposed LP method is applied to the HILN coder. Simulation results using a verification model(VM) software for real audio signals show a definite trend of significant improvement.

  • PDF

A 20 GHz low-loss dual - mode channel filter using mode matching method (모드정합법을 이용한 20GHz 저손실 이중모드 채널여파기)

  • 정근욱;이재현;유경완;강성춘
    • Journal of the Korean Institute of Telematics and Electronics D
    • /
    • v.34D no.10
    • /
    • pp.53-59
    • /
    • 1997
  • In this paper, we present a 20 GHz low-loss dual-mode channel filter designed by using mode matching method. The performance of dual-mode channel filter mainly depends on iris characteristics. Therefore the exact design of iris is the key point to get good frequncy response of the filter. MOde matching technique is widely used ot design several kinds of waveguide filters because it is simple in theory and can easily calculate the scattering matrices at the discontinuities with simple structure like iris coupled filters. Additionally the effect for finite thickness of the iris in the dual-mode cavity iflter is analyzed by te full-wave method, providing the exact filter implementation without trial and error.

  • PDF

Development of Frequency Domain Matching for Automated Mosaicking of Textureless Images (텍스쳐 정보가 없는 영상의 자동 모자이킹을 위한 주파수영역 매칭기법 개발)

  • Kim, Han-Gyeol;Kim, Jae-In;Kim, Taejung
    • Korean Journal of Remote Sensing
    • /
    • v.32 no.6
    • /
    • pp.693-701
    • /
    • 2016
  • To make a mosaicked image, we need to estimate the geometric relationship between individual images. For such estimation, we needs tiepoint information. In general, feature-based methods are used to extract tiepoints. However, in the case of textureless images, feature-based methods are hardly applicable. In this paper, we propose a frequency domain matching method for automated mosaicking of textureless images. There are three steps in the proposed method. The first step is to convert color images to grayscale images, remove noise, and extract edges. The second step is to define a Region Of Interest (ROI). The third step is to perform phase correlation between two images and select the point with best correlation as tiepoints. For experiments, we used GOCI image slots and general frame camera images. After the three steps, we produced reliable tiepoints from textureless as well as textured images. We have proved application possibility of the proposed method.