Use this URL to cite or link to this record in EThOS:
Title: Ship detection with SAR : modelling, designing and real data validation
Author: Iervolino, P.
ISNI:       0000 0004 5918 8407
Awarding Body: University of Surrey
Current Institution: University of Surrey
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
The request for maritime security and safety applications has increased in the recent past. In this scenario, Synthetic Aperture Radar (SAR) sensors are one of the most effective means thanks to their capability to get images independently from daylight and weather conditions. In the SAR ship-detection field, many algorithms have been presented in literature; however none of them has ever considered the aspects behind the interaction of the electromagnetic wave between the target and the surrounding sea. This thesis explores the electromagnetic interaction arising between the ship and the sea and, firstly, a novel model to evaluate the Radar Cross Section (RCS) backscattered from a canonical ship is derived. RCS is modelled according to Kirchhoff Approximation (KA) within the Geometric Optics (GO) solution. The probability density function relative to the double reflection contribution is derived for all polarizations and the new model is validated on SAR images showing a good match between the theoretical values and those ones measured on real SAR images. Then, a novel ship detector, based on the Generalized Likelihood Ratio Test (GLRT) where both the sea and the ship electromagnetic models are considered, is proposed. The GLRT is compared to the CFAR algorithm through Monte Carlo simulations in terms of ROCs (Receiver Operating Characteristic curves) and computational load at different bands (S, C and X). Performances are also compared through simulations with different orbital and scene parameters. The GLRT is then applied to datasets acquired from different sensors operating at different bands: the Target to Clutter Ratio (TCR) is computed and detection outcomes are compared with AIS data. Results show that the GLRT presents better ROCs and greatly improves the TCR, but its computational time is slower when compared to the CFAR algorithm. Finally, a new approach for ship-detection and ambiguities removal in LPRF (Low Pulse Repetition Frequency) SAR imagery is proposed. The method exploits the range migration pattern and is evaluated on a downsampled SAR image. The algorithm is able to reject the SAR azimuth ambiguities and can be adapted for the upcoming Maritime Mode of the future NovaSAR-S sensor.
Supervisor: Guida, R. Sponsor: Engineering and Physical Science Research Council (EPSRC) ; Surrey Satellite Technology Ltd (SSTL)
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available