Submarine Cable Tracking in Forward Looking Sonar Images

The work presented here is dedicated to the detection of submarine cable using FLS (Forward Looking Sonars). In a first part, a presentation of the dataset is realized. The data were collected in the Bay of Douarnenez (near Brest, France) with the Hector 7 ROV (Remotely Operated Vehicle) deployed from the “Pierre de Fermat” Orange Marine cable-ship, following a zig-zag trajectory so as to cross the cable while adjusting the heading. The different steps of the image processing are detailed. A preprocessing step includes successively an image normalization according to the angle and the distance, an incoherent combination of successive pings in order to improve the SNR (signal to noise ratio) and an image rotation so as to present the cable horizontally. The next processing step consists in template matching. After defining the template, it is used as a kernel to enhance cable saliency in FLS images by taking the absolute difference between correlation and convolution. The cable is then detected by a percentile-based thresholding. False alarms are removed by a tracking process, based on apparent motion in the images sequence. Plotting the cable positions in a GIS (Geographic Information System) demonstrates the ability of the proposed method to locate the cable with a reasonable accuracy.