METHODOLOGY FOR IDENTIFYING REMOTE OBJECTS USING A SET OF HETEROGENEOUS FEATURES

Development and testing of a method for using optoelectronic systems with the ability to obtain detailed and spectrophotometric information for small-sized object identification tasks. The study used the illumination characteristics of the observed objects to obtain characteristic features for comparison with base objects. In order to eliminate the influence of distorting factors, a method for identifying remote objects using a set of heterogeneous features was considered. The results were compared with a database of test object features. To assess the reliability of object identification by the algorithms of the method, a metric for comparing the variability of heterogeneous features of the observed and base objects was used. The proposed method for identifying objects was tested based on an experimental model of a hybrid optoelectronic system simulating the operation of ground-based monitoring facilities in near-Earth space. The formation of features of the observed objects is implemented by standard algorithms of the National Instruments Vision Assistant application. It has been established that the proposed technique provides a higher probability of correct identification of the observed objects based on the obtained set of heterogeneous features of low-contrast, barely noticeable remote objects. The ranges of the signal/noise ratio of objects and background, at which the results have the best indicators of the efficiency of object identification, are determined. The proposed technique allows one to evaluate the capabilities of existing optoelectronic systems for identifying remote objects using a set of heterogeneous features in difficult observation conditions, and to form the appearance of promising means of observing barely noticeable objects. The set of developed algorithms for extracting features of the observed objects allows one to identify the objects of interest, as well as to provide a higher probability of their identification. The obtained results form the basis for developing a wide range of hybrid optoelectronic ground-based means of monitoring near-Earth space.