Automatic benthic imagery recognition using a hierarchical two-stage approach
- 199 Downloads
The main objective of this work is to establish an automated classification system of seabed images. A novel two-stage approach to solving the image region classification task is presented. The first stage is based on information characterizing geometry, colour and texture of the region being analysed. Random forests and support vector machines are considered as classifiers in this work. In the second stage, additional information characterizing image regions surrounding the region being analysed is used. The reliability of decisions made in the first stage regarding the surrounding regions is taken into account when constructing a feature vector for the second stage. The proposed technique was tested in an image region recognition task including five benthic classes: red algae, sponge, sand, lithothamnium and kelp. The task was solved with the average accuracy of 90.11% using a data set consisting of 4589 image regions and the tenfold cross-validation to assess the performance. The two-stage approach allowed increasing the classification accuracy for all the five classes, more than 27% for the “difficult” to recognize “kelp” class.
KeywordsSeabed image segmentation Machine learning Supervised classification Feature extraction Two-stage classifier
The authors would like to thank Sergej Olenin and his team for allowing to use their video sequences and manual data labelling.
- 8.Galloway, W.E.: Process framework for describing the morphologic and stratigraphic evolution of deltaic depositional systems. Deltas: Models for Exploration, pp. 87–98 (1975)Google Scholar
- 10.Gleason, A.C.R., Reid, R.P., Voss, K.J.: Automated classification of underwater multispectral imagery for coral reef monitoring. OCEANS 2007, 1–8 (2007)Google Scholar
- 11.Gobi, A.F.: Towards generalized benthic species recognition and quantification using computer vision. In: OCEANS 2010, pp. 1–6. IEEE, Sydney (2010)Google Scholar
- 13.Mansoor, H., Sorayya, M., Aishah, S., Mosleh, M.A.: Automatic recognition system for some cyanobacteria using image processing techniques and ANN approach. In: International Conference on Environmental and Computer Science, vol. 19, pp. 73–78. Singapore (2011)Google Scholar
- 14.Hamilton, L.J.: Topics in acoustic seabed segmentation current practice, open software, and data fusion. In: Proceedings of International Conference Acoustics, Development, and the Environment. Fremantle (2012)Google Scholar
- 15.Jalali, S., Seekings, P.J., Tan, C., Tan, H.Z.W., Lim, J.H., Taylor, E.A.: Classification of marine organisms in underwater images using CQ-HMAX biologically inspired color approach. In: The 2013 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2013)Google Scholar
- 20.Manderson, T., Li, J., Cort’es Poza, D., Dudek, N., Meger, D., Dudek, G.: Field and Service Robotics: Results of the 10th International Conference, chap. Towards Autonomous Robotic Coral Reef Health Assessment, pp. 95–108. Springer International Publishing, Cham (2016)Google Scholar
- 21.Mosleh, M.A., Manssor, H., Malek, S., Milow, P., Salleh, A.: A preliminary study on automated freshwater algae recognition and classification system. BMC Bioinform. 13(17), 1–13 (2012)Google Scholar
- 22.Pugh, M., Tiddeman, B., Dee, H., Hughes, P.: Towards automated classification of seabed substrates in underwater video. In: 2014 ICPR Workshop on Computer Vision for Analysis of Underwater Imagery (CVAUI), pp. 9–16 (2014)Google Scholar
- 24.Šaškov, A., Dahlgren, T.G., Rzhanov, Y., Schläppy, M.L.: Comparison of manual and semi-automatic underwater imagery analyses for monitoring of benthic hard-bottom organisms at offshore renewable energy installations. Hydrobiologia 756(1), 139–153 (2014)Google Scholar
- 26.Shihavuddin, A., Gracias, N., Garcia, R., Escartin, J., Birger Pedersen, R.: Automated classification and thematic mapping of bacterial mats in the north sea. In: OCEANS-Bergen, 2013 MTS/IEEE, pp. 1–8 (2013)Google Scholar