Omani moots new ways of studying corals

By Yahya Al Salmani — MUSCAT: Dec 26: Amran Mohammed al Kamzari (Omani researcher) recently obtained his Master degree from the Department of Marine Science and Fisheries, College of Agricultural and Marine Sciences, Sultan Qaboos University (SQU). The title of his dissertation is ‘Computer-based Identification of Coral Reef Substrates Using Underwater Images’. The study is considered the first of its kind in the Middle East, says the researcher. Coral community or coral reef surveys include a variety of methods based on sampling the substrates to assess species and substrates’ abundance and cover.
A common characteristic of these methods is the time spent by experts analysing the data, either underwater as diving experts or as analysts of photographs or video transects in the laboratory.
The analysis of such underwater transects requires the identification of substrates, including corals, based on their appearance in the photographs or video frames.
The study was aimed at defining underwater image features (colour and texture) and classification algorithms to be used for a semi-automatic annotation of underwater survey videos, which may prove to be a cost-effective and time- efficient tool for reef surveys.
A series of video frames were extracted from numerous underwater survey videos of Musandam (Sultanate of Oman) coral communities.
The study provides the groundwork for developing rapid automated systems for coral reef monitoring.
The recommendations of the study are: using the double approach (feature extraction and classification), it has been recommended to encourage the development of user-friendly software for coral community monitoring.
Al Kamzari said in the study, only nine substrates were entered in the database.
It would be interesting to test the algorithms with more substrates, additional coral species and non-coral substrates such as rubbles and sponge which may be in abundance in some habitats.
It also recommends the development of a software system to include the automatic/ semi-automatic identification in a user-friendly environment.
Al Kamzari said the study provides groundwork for developing rapid automated systems for coral reef monitoring to assess the changes in live coral cover in relation to disturbances or different levels of protection.
The study proposes a method based on colour and texture features over multiple scales that outperforms classification methods and establishes a strong baseline on the dataset. It says using computer-based automatic/ semi-automatic image annotation will reduce the time spent in sampling under water.
Data acquisition underwater will reduce costs and can be undertaken by people not trained in identifying benthos communities (organisms that live in the bottom of ocean floor).