International Journal of Remote Sensing · 2007 · 101 citations · 39 references
ComputingMarine SciencesOceanographyComputer VisionCoral Reef EcologyMarine EcologyClassificationImage ClassificationObject DetectionObject RecognitionRemote Sensing Image AnalysisSonar SystemsBenthic Habitat ClassificationBenthic HabitatsHabitat MappingSide-scan SonarSide-scan Sonar ImagingSidescan Sonar MappingSonar Image ProcessingUnderwater Image Segmentation
Traditional pixel‑based classification produces a pixelized, salt‑and‑pepper representation of habitat distribution, whereas object‑based classification yields discrete habitat objects in raster or vector form. The study develops an object‑oriented side‑scan sonar classification method and evaluates whether shape, size, and texture parameters can discriminate reef from sand habitats. The method uses Edgetech 272 side‑scan sonar imagery, segments it at multiple levels, and assesses classification accuracy through visual digitization and video transects. Cross‑validation with confusion and error matrices shows the object‑based method achieves 80 % accuracy for two classes and 60 % for three classes, comparable to or slightly below digitized layers, and preserves discrete habitat patterns in an automated, repeatable vector format.
This research presents an object‐oriented technique for habitat classification at different segmentation levels based on the use of imagery from an Edgetech 272 side scan sonar. We investigate the success of object parameters such as shape and size as well as texture in discriminating reef from sand habitat. The results are evaluated using traditional digitization, based on visual assessment of the sidescan imagery, and video transects. Whereas the application of traditional pixel‐based classification results in a pixelized (salt and pepper) representation of habitat distribution, the object‐based classification technique results in habitat objects (raster or vector). The object‐oriented classification results are cross‐validated using confusion matrices in image classification software and error matrices from underwater video transects showing an overall accuracy of 80% based on two classes within the image at three segmentation levels and an overall accuracy of 60% based on three classes at two segmentation levels. This is compared with the digitized layer accuracy of 81% for two classes and 72% for three classes, and this demonstrates the successful application of object‐oriented methods for habitat mapping. This technique retains spatially discrete habitat pattern information in a classified vector shape file with methods that are automated, repeatable, objective, and capable of processing many sidescan records in a more efficient manner.
39
Photogrammetric Engineering and Remote Sensing
John Nicol, B. King · UCL Discovery (University College London) · 2007 · 9.4K citations