K.K. Vasiliev - Dr. Sc. (Eng.), Professor, Head of Department «Telecommunications», Ulyanovsk State Technical University E-mail: vkk@ulstu.ru V.E. Dementiev - Ph. D. (Eng.), Associate Professor, Department «Telecommunications», Ulyanovsk State Technical University E-mail: vitawed@mail.ru N.A. Andriyanov - Post-graduate Student, Department «Telecommunications», Ulyanovsk State Technical University E-mail: nikita-and-nov@mail.ru

One of the main problems in statistical radio engineering which has always been considered is the problem of signal detection. At the same time, there are a number of such tasks where a noise is correlated. A striking example is the radar purposes detection when there is interference from rough sea surface, or the detection of a bright object in the presence of heterogeneous background in optical or radiometric monitoring systems of the earth\'s surface. It is known that the construction of the signal detection algorithms are often based on selecting a model. Thus the development of new models that can improve the quality of the description of real signals, and consequently, the accuracy of processing entails the synthesis and study of signal detection algorithms on the background of these models. Thus, the proposed model with a complex structure should be investigated with respect to the signal detection problem. This article discusses one of the most urgent tasks of image processing. It is the problem of signal detection. The aim of the authors is to investigate the effectiveness of detection in the case of doubly stochastic models of random fields, and in the case of AR models. The signal is detected on the background of correlated random field with a complex structure. The authors use statistical modeling for signal detection in random processes. The well-known signal is added to random process in well-known area. According to the results of statistical modeling it is conducted to compare the efficacy of detection on the basis of two types of models. In addition, there is the investigation of the detection algorithm for different parameters of the models. The dependence of correct detection probability of the extended signal on the signal level and for different values of internal correlation coefficients. Advanced statistical modeling was carried out for the case of the detection of the signal point. The authors noted that the results obtained for the correct detection probability of the point signal detection and the extended signal with length of 4 counts is close enough. The ratio of the length of the signal to the all sequence length is approximately 0.001. Analysis of the results shows that the detection efficiency will be higher for the doubly stochastic models. In addition, the detection efficiency also depends on the model parameters, and it is higher, when the correlation coefficient is higher. At the same time a consequent increase of the internal correlation coefficient leads to the approximation of a doubly stochastic model to AR model. This is a reason of smaller gain in detection. However, even with the maximum correlation, when coefficient is very close to 1, algorithm based on a doubly stochastic model provides better results.

 

1. SHakhtarin B.I. Obnaruzhenie signalov: Ucheb. posobie dlja vuzov. M.: Gorjachaja linija-Telekom. 2014. 526 s.
2. SHalygin A.S., Palagin JU.A. Prikladnye metody statisticheskogo modelirovanija. L.: Mashinostroenie. 1986. 320 s.
3. Vasilev K.K., Krasheninnikov V.R. Statisticheskijj analiz izobrazhenijj. Uljanovsk: UlGTU. 2015. 214 s.
4. Vasilev K.K., Tashlinskijj A.G., Krasheninnikov V.R., Statisticheskijj analiz posledovatelnostejj mnogomernykh izobrazhenijj // Naukoemkie tekhnologii. 2013. № 5. S. 5−11.
5. Vasilev K.K., Dementev V.E., Andrijanov N.A. Ocenivanie parametrov dvazhdy stokhasticheskikh sluchajjnykh polejj // Radiotekhnika. 2014. № 7. S. 103−106.
6. Vasil\'ev K.K., Dement\'ev V.E., Andriyanov N.A. Application of mixed models for solving the problem on restoring and estimating image parameters // Pattern Recognition and Image Analysis. January 2016. V. 26(1). P. 240−247.