Search academic texts

Information × Registration Number 0218U001226, 0116U004638 , R & D reports Title Automation of recognition and classification of suspicious objects for video data from unmanned aircraft cameras popup.stage_title Head Prystavka Phylyp, Registration Date 26-01-2018 Organization National Aviation Univercity popup.description2 The research object is the search and classification of objects on the video from the unmanned aerial vehicle camera (UAV). The subject of the study is the methods of recognition and classification of suspicious objects according to the data from the suspicious load cells UAV. The purpose of the work (in accordance with the specifications) is to improve the well-known, development of new methods for recognition and classification of suspicious objects based on data from the unmanned aerial vehicle suspicious loading cameras (UAV), as well as the creation of experimental prototype software of the automated recognition and classification system on their basis target objects for video data from UAV suspicious load cells. The main scientific results are as follows: 1) the new linear operators of the partial derivatives of the first and second order functions of two variables are the first obtained, based on the model of the image in the form of a linear combination of B-splines of the second and higher orders, close to the interpolation average, which is the basis of the mathematical apparatus for the results given further; 2) the method of large-scale analysis of a digital image is improved, proceeding from the model of the image in the form of a linear combination of B-splines, which provides higher performance when implemented in software compared with the approach based on the use of two-dimensional gaussians; 3) at first time, a new method for recognizing and classifying suspicious objects on a digital image and video was developed for the first time, based on the model of the image in the form of a linear combination of B-splines, which, unlike the known SIFT method and some of its modifications, has no patent constraints; 4) for the first time the information technology of adaptive search, classification and recognition in the streaming video from UAV cameras has been developed; objects similar to the target, which can be the core of automated search systems for suspicious objects for problems of different character; 5) software is developed that implements the information technology of adaptive search, classification and recognition of objects with self-study, and which is integrated into the developed prototype of the software-hardware complex of search of suspicious objects. The novelty of the results obtained is that among the analogues that are not secret, in particular, presented at the XIV International Specialized Exhibition "Weapons and Security - 2017" on October 10-13, 2017, Kyiv, the function of the automated search of suspicious objects for no conditions for the absence of reference samples. The subsystem of the determination of the coordinates of the identified objects gives less accuracy in comparison with the declared (but not certified) accuracy of analogues (development of the State Enterprise "Ukroboronprom", enterprises of the NGO "League of Defense Enterprises of Ukraine", etc.). The results of the research work are used by the State Enterprise "Orizon Navigation", which is an appropriate implementation act, in particular: information technology for the recognition and classification of objects from the UAV cameras and the synchronization of the frame with the data of the satellite navigation system in the automatic mode; Post-processing information technology based on histogram analysis and modified recognition method based on the analysis of special points based on the spline-model of the image. Product Description popup.authors Ільницька Світлана Іванівна Артамонова Наталія Георгіевна Білик Артем Юрійович Балацький Сергій Сергійович Бурий Павло Андрійович Вишнякова Євгенія Вікторівна Дегтярев Виктор Васильович Задонцев Юрій Вікторович Захарчук Тамара Миколаівна. Кондратюк Василь Миколайович Кондратюк Маріна Василівна Курочкін Віктор Михайлович Куценко Олександр Вікторович Піскунов Олексій Германович Пантелеев Вячеслав Михайлович Приставка Пилип Олександрович Прокопенко Ігор Григорович Прокопенко Костянтин Ігорович Тиводар Оксана Василівна Трикоз Валерій Павлович Чеканов Євген Миколайович Чирков Артем Валерійович Чолишкіна Ольга Генадіївна Шевченко Андрей Костянтинович Юрчук Ірина Аркадіївна Ясенко Сергій Анатолійович popup.nrat_date 2020-04-02 Close