🤖 AI Summary
This work addresses the limitations of existing oriented bounding box (OBB) to horizontal bounding box (HBB) conversion methods in remote sensing ship detection, which often introduce excessive background clutter or lose critical target details, thereby degrading detection accuracy. To mitigate this issue, the study proposes a shape-aware HBB generation strategy that, for the first time, incorporates geometric characteristics of ships—including contour shape, fill ratio, and OBB orientation—into the conversion process to produce tighter, more target-aligned axis-aligned bounding boxes. Experimental results demonstrate that the proposed method significantly improves bounding box compactness and detection robustness compared to baseline approaches such as Outer HBB, Area Equivalent HBB, and GBB Marginalized HBB, effectively preserving essential target regions.
📝 Abstract
Accurate object detection in aerial and satellite imagery is dependent upon the bounding box representation. This is especially true for spatially oriented objects such as ships or aircrafts. Oriented Bounding Boxes (OBB) have a tighter fit and more robust non-max suppression compared to Horizontal Bounding Boxes (HBB), any current post-processing conversion from OBB to HBB either introduces excess empty and background space or removes data from the detection. This paper introduces a novel approach for a shape-aware OBB-to-HBB conversion for ship detection in remote sensing imagery. It leverages hull shape, hull fullness, and the bounding box orientation to produce a tighter axis-aligned HBB representation. The proposed method is benchmarked against three baselines methods for OBBto-HBB conversion, Outer HBB which uses minimum and maximum, Area Equivalent HBB and GBB Marginalized HBB.