Image-Based Coral Reef Classification and Thematic Mapping

Abstract

This paper presents a novel image classification scheme for benthic coral reef images that can be applied to both single image and composite mosaic datasets. The proposed method can be configured to the characteristics (e.g., the size of the dataset, number of classes, resolution of the samples, color information availability, class types, etc.) of individual datasets. The proposed method uses completed local binary pattern (CLBP), grey level co-occurrence matrix (GLCM), Gabor filter response, and opponent angle and hue channel color histograms as feature descriptors. For classification, either knearest neighbor (KNN), neural network (NN), support vector machine (SVM) or probability density weighted mean distance (PDWMD) is used. The combination of features and classifiers that attains the best results is presented together with the guidelines for selection. The accuracy and efficiency of our proposed method are compared with other state-of-the-art techniques using three benthic and three texture datasets. The proposed method achieves the highest overall classification accuracy of any of the tested methods and has moderate execution time. Finally, the proposed classification scheme is applied to a large-scale image mosaic of the Red Sea to create a completely classified thematic map of the reef benthos.

In situ Analysis of Coral Recruits Using Fluorescence Imaging

“…Small-scale macro photography has also been used for imaging coral recruits (Edmunds et al, 1998). In the larger scale, using photographic surveys of the seabed is becoming a common method for mapping reef communities (Shihavuddin et al, 2013;Cardini et al, 2015). Advances in computer vision have enabled the processing and automated annotation of these large-scale coral reef surveys (Stokes and Deane, 2009;Beijbom et al, 2012Beijbom et al, , 2016, offering a powerful alternative to manually annotated visual surveys (Singh et al, 2004;Beijbom et al, 2015).…”

Ground-Level Classification of a Coral Reef Using a Hyperspectral Camera

“…While this is spectrally true, the analysis misses an important detail relevant to temporal changes on the reef-coral mortality. Because spectral separation is not possible here, future attempts should consider object or pattern recognition protocols [55,56] as a secondary classification stage.…”

“…These are spectrally similar but represent an important shift in substrate change. The solution probably lays in the application of object recognition tools that would separate growth patterns (e.g., [55,56]). A recent review by Huang et al [63] systematically compares a variety of classifiers using combinations of spectral and semantic features datasets.…”

“…Other classification methods relevant to hyperspectral imagery that are worth testing include random forest [25,62] and Support Vector Machine (SVM) (e.g., [13,56,63]). Random forest is reported mostly with regards to increasing accuracy when spectral resolution increase [13,64] and combined with the uses of indices [62].…”

Autonomous Coral Reef Survey in Support of Remote Sensing

“…in situ measurements of benthic reflectance using diver-operated spectrometers have, for example, highlighted the spectral differences between various coral reef components and aided assessments of remote sensing systems for coral reef mapping (Hochberg and Atkinson, 2000;Hochberg et al, 2003;Kutser et al, 2003), but field operations are laborious and inefficient for surveying purposes. In situ photogrammetric and spectral imaging approaches to mapping coral reef structure have been reported using diver-operated and towed systems (e.g., Gleason et al, 2007;Lirman et al, 2007;Shihavuddin et al, 2013). While these approaches have yielded high-quality photo mosaics and thematic images of benthic features that can be used to validate remote sensing products, such as benthic cover, the data are generally qualitative from a radiometric perspective and of limited use in developing and testing shallow-water light models.…”