Publication: Gemi akustik izlerinin otomatik sınıflandırılması
Abstract
Sualtı akustik sinyallerin sınıflandırılmasında SONAR sistemleri ile elde edilen bilgiler kullanılarak operatörler tarafından gerçekleştirilmektedir. Askeri alanda savaş durumlarında dost/ düşman hedefleri tespit/ teşhis ve kimliklendirilmesinde SONAR operatörüne destek olması için otomatik sınıflandırma algoritmaları kullanılması gerekmektedir. Bu tez çalışmasında sualtı akustik sinyaller kullanılarak su üstünde yer alan hedef ya da nesnelerin otomatik olarak sınıflandırılması gerçekleştirilmiştir. Su üstü ve sualtı platformların sualtına yayılan gürültülerinde baskın olanı pervane kavitasyonu sonucu ortaya çıkan geniş bant gürültüleridir. Kavitasyon gürültüsünü kullanılarak hedef tespit ve tanıması için Gürültüde Genlik Modülasyon Tespit (DEMON, ing. Detection Envelope Modulation on Noise) algoritması geliştirilmiştir. Literatürde yer alan DEMON algoritmaları geliştirilerek elde edilen yeni DEMON algoritması ile sınıflandırma performans analizleri ve katkıları incelenmiştir. Bu tez çalışması kapsamında önerilen DEMON algoritması açık kaynak veri setleri kullanılarak literatürde yer alan diğer çalışmalar ile performans karşılaştırılması yapılmıştır. Literatürde yer alan veri setlerine örnek oluşturması maksadıyla benzer yöntemler kullanılarak sahadan veriler toplanmıştır. Toplanan veri ile elde edilen algoritmanın doğrulandığı gözlemlenmiştir.
Underwater acoustic signals are classified by operators using information obtained from SONAR systems. Automatic classification algorithms should be used to assist the SONAR classify in detecting/ diagnosing and identifying friendly/ hostile targets in combat situations in the military domain. In this study, the automatic classification of targets or objects above water was performed using underwater acoustic signals. The dominant noise among the underwater radiated noise of surface and underwater platforms is the broadband noise that occurs as a result of propeller cavitation. Detection Envelope Modulation on Noise (DEMON) algorithm has been developed for target classific and recognition using cavitation noise. With the new DEMON algorithm obtained by developing DEMON algorithms in the classifica, classification performance analyzes and contributions are examined. The DEMON algorithm proposed within the scope of this thesis study was compared with other studies in the classifica by using open source datasets. The data obtained with the proposed In order to set an example for the data sets in the literature, data were collected from the field using similar methods. It was observed that the algorithm obtained with the collected data was verified.
Underwater acoustic signals are classified by operators using information obtained from SONAR systems. Automatic classification algorithms should be used to assist the SONAR classify in detecting/ diagnosing and identifying friendly/ hostile targets in combat situations in the military domain. In this study, the automatic classification of targets or objects above water was performed using underwater acoustic signals. The dominant noise among the underwater radiated noise of surface and underwater platforms is the broadband noise that occurs as a result of propeller cavitation. Detection Envelope Modulation on Noise (DEMON) algorithm has been developed for target classific and recognition using cavitation noise. With the new DEMON algorithm obtained by developing DEMON algorithms in the classifica, classification performance analyzes and contributions are examined. The DEMON algorithm proposed within the scope of this thesis study was compared with other studies in the classifica by using open source datasets. The data obtained with the proposed In order to set an example for the data sets in the literature, data were collected from the field using similar methods. It was observed that the algorithm obtained with the collected data was verified.