Research On Passive Sonar Display Optimization Technology

The sonar display technology includes post-processing, image processing as well as thedesigning of display screen which are related to sonar display. It improves the detectionperformance of sonar by analysing and processing the output of signal processing module anddisplaying as much as possible of the information on the screen. In order to enhance thedetection performance of sonar, this paper focused on passive sonar display optimizationtechnology, especially the background equalization technology and line spectrum detectiontechnology based on the LOFAR gram.First of all, the paper discussed the common display modes of passive sonar andpresented the display formats combined with commonly used signal processing techniques.Then it researched the ‘Trace to Trace Correlation’ effect and linear/non-linear gray level setpattern.In accordance with the uneven characteristics of background noise in broadband energydetector, two kinds of background equalization algorithm were studied by computersimulation, including OTA(Order Truncate Average) algorithm and a background equalizationalgorithm based on beam characteristics. The former was able to remove a wide range ofnoise by combining the threshold comparing and median filtering with the weak point ofbeing sensitive to the parameters. The latter could suppress strong interference while retainthe weak signal with no need to set the parameters.At last, the line spectrum detection technology based on the LOFAR gram wasthoroughly studied, including the line spectrum detection algorithms based on secondaryjudgment and optimal path. The former could present a clear image in real time with a lowSNR via simulating the dual threshold detection process performed by the sonarman whenobserving the LOFAR gram. However, it had a high requirement of the threshold setting andfrequency stability. The latter utilized the optimal path method with a cost function built byamplitude, frequency difference and intermittent points. This algorithm applied to linespectrum with a constant or unconstant frequency, but its performance was affected by thelength of the observation window.