XXII ISPRS Congress 2012: Technical Commission VIII

Shortis, M.; Shimoda, H.; Cho, K.

Bibliographic data

Multivolume work

Persistent identifier:: 1663813779

Title:: XXII ISPRS Congress 2012

Sub title:: Melbourne, Australia, 25 August-1 September 2012

Year of publication:: 2013

Place of publication:: Red Hook, NY

Publisher of the original:: Curran Associates, Inc.

Identifier (digital):: 1663813779

Language:: English

Additional Notes:: Kongress-Thema: Imaging a sustainable future

Corporations:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Adapter:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Founder of work:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Other corporate:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Document type:: Multivolume work

Volume

Persistent identifier:: 1663822514

Title:: Technical Commission VIII

Scope:: 590 Seiten

Year of publication:: 2014

Place of publication:: Red Hook, NY

Publisher of the original:: Curran Associates, Inc.

Identifier (digital):: 1663822514

Illustration:: Illustrationen, Diagramme

Signature of the source:: ZS 312(39,B8)

Language:: English

Additional Notes:: Erscheinungsdatum des Originals ist ermittelt.; Literaturangaben

Usage licence:: Attribution 4.0 International (CC BY 4.0)

Editor:: Shortis, M.; Shimoda, H.; Cho, K.

Corporations:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Adapter:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Founder of work:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Other corporate:: International Society for Photogrammetry and Remote Sensing, Congress, 22., 2012, Melbourne; International Society for Photogrammetry and Remote Sensing

Publisher of the digital copy:: Technische Informationsbibliothek Hannover

Place of publication of the digital copy:: Hannover

Year of publication of the original:: 2019

Document type:: Volume

Collection:: Earth sciences

Chapter

Title:: [VIII/9: Oceans]

Document type:: Multivolume work

Structure type:: Chapter

Chapter

Title:: EXTRACTION OF BENTHIC COVER INFORMATION FROM VIDEO TOWS AND PHOTOGRAPHS USING OBJECT-BASED IMAGE ANALYSIS M. T. L. Estomata, A. C. Blanco, K. Nadaoka, E. C. M. Tomoling

Document type:: Multivolume work

Structure type:: Chapter

require such information (Marcos, et al., 2008). Since precise data will be readily available, reef processes can be more realistically studied as well as serve as a quick aid in developing improved management strategies (Scopélitis, et al., 2010). Established Marine Protected Areas (MPAs) increase the importance of acquiring data from monitoring to determine if they are able to achieve their management goals (Hill & Wilkinson, 2004). MPAs in the Philippines will then be able to develop better management schemes, especially during these times of climate change when marine habitats are very vulnerable. 1.3 Scope and limitations The main scope and focus of this research is to use object-based classification on underwater photos to extract cover information. Hence, the study mainly investigates the performance of OBIA vis-a-vis typical pixel-based spectral classifiers. It deals with the creation of rule sets necessary to obtain high classification accuracy using the OBIA approach. The chosen 50-meter transect was approximately 10 meters deep and the video used for producing the benthic cover map was taken at a 6-m depth, about 3-4 meters from the reef surface. Due to this distance of the camera from the reef and the absence of rocks, sea grass and other benthos, only a general classification was produced — coral, sand and rubble. Since a MRU nor a small bubble level were not available to ensure the vertical orientation of the underwater camera used during the fieldwork, georeferencing from one video snapshot to another lead to high root mean square error (RMSE), thus an uncontroled mosaic (ie. snapshots that were subjectively stitched together using Adobe Photoshop) was used as an alternative to create a panorama of the transect. An external motion sensor input for the MBES is the most important add-on because it compensates for roll, pitch and heave of the beams in real time but it was not available during the data gathering. Since corrections were not available and applied, the accuracy of the bathymetric data (vertical measurements) was not reliable enough for the underwater video snapshots to be georeferenced to it. Due to this, classification between elevated and non-elevated features could not be accurately performed. With the available MBES data, the uncontrolled mosaic was given only an approximate correction for scale and rotation, but its geographic location (horizontal measurements), through georeferencing to the relative depth of the MBES data, is accurate up to the centimeter level. A water level logger, which would record the rise and fall of tide, was also not available during the data gathering, thus a tide predictor program (WXTide) was used to supply tide information during the fieldwork. The data supplied by the program was used to generate corrections for the bathymetric data gathered by the MBES. 1.4 References Acquisition and classification of underwater videos and photographs. Many studies have used video surveys and photographs to capture images of coral reefs, which were then used for classification. A study evaluated the different survey techniques used to validate maps derived from remotely sensed images and was able to conclude that among all survey techniques they evaluated, the best choice, if resources were unlimited and expertise was available, was the photographic transect, which was processed using the 1024 point analysis (Roelfsema, et al, 2006). Based on another study of these remotely sensed images, they said that manual delineation with field validation resulted to the highest accuracy (Scopélitis, et al, 2010). A research (Kaeli, et al., 2005) focused on deep water corals, which were present at around 30 — 100 meters, depths which are not reachable by SCUBA diving. Thus, the SeaBED Autonomous Underwater Vehicle (AUV) took the images, which were then analyzed using the existing random point method. They identified the Montastrea annularis complex, which was a dominant coral in their area and had a smooth texture. They also calculated the percent coverage of the coral upon classification. Another study (Marcos, et al., 2008) which used underwater videos recorded them at a near-reef distance, where the video was approximately 30cm from the reef surface. The video was divided into 625 sub-images, which were then used to train the test set to classify between living (live coral and algae) and non-living (dead coral, sand and rubble). The classifier used was based on the Bayesian theory called linear discriminant analysis (LDA) and both color and texture features were inputs to the classifier system. 79% was the overall success rate of the near-reef videos and at shallower depths, like 3 meters, a recognition rate of 85% was acquired for the living and 75% for nonliving. The neural network approach was also explored and used in a research (Marcos, et al., 2005).They captured the video at a constant range of 15-30 cm from the reef surface and the images that they classified were close-up of coral reefs. They used of a "feed-forward back-propagation neural network classification" to classify the images into three benthic categories: live coral, dead coral and sand. Color and texture features were inputs to the network and they were able to obtain 86.5% success rate for test images that were not included in the training set and 79.7% recognition rate for the same set of images. A research (Scopélitis, et al., 2010) focused on satellite images and used three different classification methods — expertise- based, pixel-based and object-based. The expertise-based classification used ArcGIS 9.2 ® to apply visual analysis and manual delineation of polygons. These maps were the most accurate, based on the observation that sites for field validation were all classified correctly, thus they were used as reference so as to provide assessment of pseudo-error on the validity of the maps that were produced using pixel and object-based. All pixels of the reference image map were cross tabulated with the maps produced from object-based and pixel-based methods. The pixel-based classification was accomplished using ENVI 4.4®’s maximum likelihood classification algorithm, which was trained using the sites which were known to be coral communities. Object-based classification used the software Definiens 7.0 ®, which divided the image to objects, which were then aggregated based on rules made by the user. The results of this research discovered that overall agreement of the object-based map with the expertise-based map was better than the pixel-based. Exploring the object-based classification method’s capability in mapping benthic cover may lead to improved results such as higher accuracies and faster result acquisition. Precision of multi-beam echo sounder. The precision of the use of multibeam echo sounding with high resolution, accompanied by positioning with high-accuracy Was investigated in a paper (Ernstsen, et al., 2006). Repetitive bathymetric data of a shipwreck was measured using a high resolu long r: annua seven the M vertice precis compa confid respec MBES the st throug meter) Integr using OBIA were | disco: compl Rangi descri pixel-l unsup hetero object to prc sugge: classi accura With f reliabl exploi acquit higher resolu MBE: be us accurz the bz reliab 21:8 Puertc provir Isla \ 1980s numb Garde Puert meter: Vic i Wikip:

Access restriction

Copyright

fullscreen: Technical Commission VIII (B8)

Multivolume work

Volume

Chapter

Chapter

Table of contents

Cite and reuse

Volume

Chapter

Image

Image fragment

Citation links

Citation recommendation