Influence of multi-look processing to SAR ship detection

Single look complex data collected by SIR-C SAR are used to generate images of different looks by multi-look processing. The equivalent number of looks (ENL), spatial resolution, and ship-sea contrast of these images are analyzed. It shows that on the one hand, multi-look processing increases ENL hence improves image quality which is good for ship detection; on the other hand, multi-look processing also decreases spatial resolution and ship-sea contrast which goes ill with ship detection. Two typical ship detection algorithms such as adaptive threshold algorithm (ATA) and K-distribution based algorithm (KBA) are used to detect ship targets from these multi-look processed images. Case studies show that the figures of merit (FOMs) for 3-look processed images using ATA algorithm and FOMs for 2-look processed images using KBA algorithm are relatively larger. This means the most appropriate choice of multi-look processing of SAR images for ship detection is 2 or 3-look processing.