A novel method of dynamic monitoring and parameter estimation for rock-fall based on multichannel SAR

Abstract

Rock-fall incidents have the characteristics of multiple, sudden and random, which threaten the surrounding buildings and traffic lines. This paper studied the monitoring and prediction of rock falls with a new synthetic aperture radar (SAR) technology. One of the most important factors in rock-fall target detection is the signal-to-clutter-plus noise ratio (SCNR) after clutter suppression, which should be maximized before rock-fall target detection. Through analysing the remainder characteristics of rock-fall targets after clutter suppression, a method of removing the quadratic frequency-modulated (FM) component associated with the azimuth velocity and along-track velocity of a rock-fall target is proposed. After removing the quadratic FM component using dechirp technology, focus is placed on the remainder signal of a rock-fall target in a Doppler domain image, and the SCNR of the signal is maximized. This method provides effective rock-fall target detection. To resolve the contradiction between the computational accuracy and the accuracy of the along-track velocity, this paper adopts gradual approach technology. Finally, by analysing the focused signal characteristics, a method of rock-fall target parameter estimation is proposed. The effectiveness of the presented method is demonstrated by both theoretic analysis and simulations.