Labelled rectangular Regions of Interest (ROI) in underwater images should be automatically detected for underwater inspection. The images contain typical underwater scenes consisting of basically three parts: Background (water); Underwater constructions (e.g., cylindrical piles); Surfaces with and without biological coverage (e.g., pocks).

The aim is the development of an automatic bounded region classifier that is at least able to distinguish between background, construction, and construction + coverage classes. The challenge is the low and varying image quality that typically appears in North- and East-sea underwater imaging. The images, typically recorded by a human diver or an AUV, pose low contrast, varying illumination conditions and colours, different viewing angles and spatial orientation and scale, overlaid by mud and bubbles (e.g., from the air supply), and optical focus issues.

We propose and evaluate a hybrid approach with segmented classification using small-scaled CNN classifiers (with less than 1000 hyper parameters) and a reconstruction of labelled ROIs by using an iterative mean and expandable bounding box algorithm. The iterative bounding box algorithm combined with bounding box overlap checking suppress spurious wrong segment classifications and represent the best and most accurate matching ROI for a specific classification label, e.g., surfaces with pocks coverage. The overall classification accuracy (true-positive classification) with respect to a single segments is about 70%, but with respect to the iteratively expanded ROI bounding boxes it is about 90%.