Improving Feature Perception in Sonar Displays by Contrast Normalisation andEnhancement

摘要

In seabed surveillance, the nonlinearity of the sonar display, together withvariations in brightness and contrast arising from variability in the stored sonar data, will have a direct effect on the visual discrimination of seabed features. The loss of information at the man-machine interface is due to environmental and instrumentation effects, and can be partially corrected for in real-time (i.e. at video frame rates) if the data is digitised and available for computer processing. Methods are described for normalising 2-D sonar data and for correcting for luminance compression in the electronic display. Experimental results for sidescan sonar indicate that statistical normalisation and display calibration by gamma-compensation provide a standardised presentation of data. The procedure improves both the visibility of seabed features and the reliability of visual search during routine surveillance.