Small bottom object density analysis from side scan sonar data by a mathematical morphology detector

Abstract

This paper describes a system for estimating the local density of small objects on the sea floor by exploiting a robust, non-parametric detector for high resolution images acquired by side scan sonar sensors. Low grazing angle target images are characterised by an area of strong intensity, the high-light, close to an area, the shadow, at the sensor noise level. The detector makes use of mathematical morphology to detect the highlight and the shadow areas within the image, and a fusion scheme to reduce the false alarms due to the sea floor disturbance by declaring a target acquired if an highlight is close to a shadow. Results on data sets collected in the Ligurian and Baltic Sea are reported and discussed