*4.2. Qualitative Results*

Figure 12 shows the visualization results on the LS-SSDD-v1.0 as an example. As we can see, Lite-YOLOv5 can carry out accurate SAR ship detection even in difficult conditions (i.e., larger scene image, multi-scale ships, and different aspect ratios of ships).

Figure 13 shows the detection results of different methods under complicated scenarios (i.e., offshore scenes of strong speckle noise and inshore scenes). Note that we only chose some lightweight models for fair comparison.

From Figure 13, one can conclude the following:


always produce the missed alarms caused by them. Taking the fourth line of images as an example, there were two missed detections of RetinaNet and two missed detections of YOLOv5, which are both more than that of Lite-YOLOv5 (only one missed ship).

3. Lite-YOLOv5 can offer an advanced on-board ship detection performance compared with other state-of-the-art methods.

**Figure 12.** The qualitative SAR ship detection results of Lite-YOLOv5. A score threshold of 0.25 is used for display. Best viewed in zoom in.

**Figure 13.** *Cont*.

(**a**) (**b**) (**c**) (**d**)

**Figure 13.** The qualitative SAR ship detection results of different methods: (**a**) ground truth; (**b**) RetinaNet; (**c**) YOLOv5; (**d**) Lite-YOLOv5. The ground truths are marked by green boxes. Prediction results are marked by yellow boxes with confidence scores. The false alarms are marked by orange ellipses. The missed detections are marked by red ellipses. GT means the number of ground truths.FN means the number of missed detections. FP means the number of false alarms.
