6533b83afe1ef96bd12a714e

RESEARCH PRODUCT

Identification of Objects Based on Generalized Amplitude-Phase Images Statistical Models

Piotr KoczurSławomir StemplewskiViktor Vlasenko

subject

0209 industrial biotechnologyamplitude-phase imagesMatching (graph theory)Plane (geometry)Computer scienceIsotropydynamic object identificationStatistical model02 engineering and technologyResidualDomain (mathematical analysis)law.invention020901 industrial engineering & automationlawAdjacency listCartesian coordinate systemgeneralized Hilbert transformsAnisotropyAlgorithm

description

The article presents the dynamical objects identification technology based on statistical models of amplitude-phase images (APIm) – multidimensional data arrays (semantic models) and statistical correlation analysis methods using the generalized discrete Hilbert transforms (DHT) – 2D Hilbert (Foucault) isotropic (HTI), anisotropic (HTA) and total transforms – AP-analysis (APA) to calculate the APIm. The identified objects are modeled with 3D airplanes templates rotated in space around the center of Cartesian coordinate system. The DHT domain system of coordinates displaying the plane projections (2D flat images) remains to be space-invariant. That causes the anisotropic properties of APIm and makes possible the tested objects effective matching to rotated templates and identification of shapes at DHT domains. As additional method for objects matching accuracy increasing the difference (residual) relative shifted phase (DRSP-) images templates are proposed. The hierarchical system for identification is based on correlation analysis and decision making on semantic models – sets of AP-histograms (adjacency arrays), DRSP-images, APIm with specified angles shifts.

yearjournalcountryeditionlanguage
2017-09-07
https://doi.org/10.1007/978-3-319-67229-8_6