A methodology correlating digital imagery with concentration via color feature extraction

• Main Authors: Mulingbayan, Isa Mae C., Sanchez, Azzedine Erika C.
• Format: text
• Language: English
• Published: Animo Repository 2013
• Subjects: Digital images; Extraction (Chemistry)
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/6664
• Institution: De La Salle University

The study devised a method of digital imagery and color feature extraction to serve as a foundation for future application in algal bloom detection. Through the extraction of the Hue-Saturation pair and Count values of each image, its correlation with the concentration of the solution is obtained. The study is focused on aerial imagery of green dye solutions with varying concentrations to represent algal blooms. Data collected show that a solution with a low count value is homogeneous and a solution with high saturation is more concentrated. A setup that provides an appropriate environment for optimum feature extraction consists of a clear matter container with a white base, a camera with at least 10 megapixels, an ISO setting of 400 or lower and a high aperture. In addition to these settings, the image must be shot in ambient light. In order to obtain the Hue and Saturation values, the images were processed in a Matlab environment that feature-extracted an image per pixel and obtained the Hue, Saturation and Count Values.