Features extraction from carbonate modern platforms satellite imagery using Computer Vision
share
Four selective, global – scale zones for carbonate modern platform sites were selected on exploring the automatic metrics extraction. The results among the relationship between modern carbonate geometric morphology and the surrounding oceanographic conditions can potentially be used to automate extracting meaningful information for geographical object recognition reservoir modeling techniques, and remote sensing using a large amount of imagery data. Define geometric features (metrics) from satellite images is a very arduous task and time-consuming work when is manually prepared. Nevertheless, digital image recognition and automatic geometric features detection is a challenging problem due to the varying lighting, orientation, and background of the target object, particularly when analyzing raw images in RGB format. In this work, an algorithm programmed in Python is presented with the objective of estimate automatically the geometric properties of a set of modern carbonate platforms located in Tun Sakaran Marine Park (Sabah, Malaysia), Kepuluan Seribu (Java Sea), Spratly Islands (China Sea) and Maldives (Indian Ocean). Firstly, a Python code load massive satellite imagery from a specific folder RGB format, then each raw coral reef image is resized, converted from RGB band to gray, smoothed, and binarized using Open Computer Vision Library available in Python 3.0. The coral reef edge information contains very prominent geometric attributes that characterize their behavior, thus morphological changes were applied to define the contour of the carbonate platform. Furthermore, a structural analysis and shape descriptors were made in a set of images in order to numerically calculate the characteristics of the carbonate platform. A total of 27 satellite images were processed by the algorithm successfully at the same time, only two images were not segmented correctly because of the illumination and intensity of the predominant colors, especially blue color. Finally, this dataset was exported to Microsoft Excel spreadsheets and CSV format respectively.
READ FULL TEXT
