Aerospace Instrument-Making Annotation << Back THE ROLE OF INFORMATIONAL SIGNS IN DETECTION, DIFFERENCES AND IDENTIFICATION OF OBJECTS Yu.S. Sagdullaev, S.D. Kovin This article discusses the detection, discrimination and identifi cation of objects in multi-component optical images (OI). Basic information features of objects – spectral-energy characteristics (EPAS), their spatial and additional (dynamic) characteristics in different spectral images. It is shown that for each object there is an information vector that describes it in a particular reception area of radiant (luminous) fl ux on the basis of these characteristics. To formalize the process of detection, discrimination and identifi cation (recognition) of objects, any object of {N}, considered as some multi-level source of information in the form of the sum of the individual components. Shows the achieved amount of information that can contain an object in all different spectral region of interest. The result of the detection, discrimination and identifi cation of object in each case is probabilistic in nature and is determined by several factors, of which the principal is the same source of information in the memory of the recognition system (RS). The total number of a priori information for the object in RS, also represented as the sum of the individual components for all areas of registration. On the basis of a specifi c ratio between the amount of information contained in the RO and RS in their interaction, recorded General condition for recognition of an object. Based on the use of a plurality of registration zones of radiant (luminous) flux in the selected spectral plot, with the formation of signals of different spectral images, we can speak about the spectral portrait of the object. He will represent the distribution of BOT objects (signals, images) in a certain region of the spectrum. Given the spectral-energy characteristics of a single arbitrary object classes, show that the use of multiple registration zones of radiant (luminous) flux of {M}, to allow to share among themselves almost all the feature classes from the set {N} and implement: 1) spectral selection of objects of different classes by the maximum value of their refl ectivity in a particular reception area of the optical spectrum; 2) enforcement operations automatically detect, distinguish and identify the objects on the distribution of their spectral characteristics in a given wavelength interval. Key words: optical imaging, detection, discrimination and identification of objects, information signs, reception area radiant flux, spectral portrait of objects. Contacts: E-mail: sagdul@mniti.ru Pp. 34-43.