Justification of Image Format Structure for Solving Pattern Recognition Problems in Environmental Monitoring

Abstract

Introduction. Environmental monitoring of marine areas and timely detection of the results of man-made disasters on the water surface, including identification of oil contaminated areas, represents an urgent task. To facilitate its solution, marine areas are controlled using space and airborne means. However, the volume of data subject to control is constantly growing. Therefore, the problem can be solved by selecting only those photo and video materials with detected traces of oil spills and other man-made disasters.Aim. To develop an approach to automatic selection of images obtained from visual control systems, providing only relevant images.Materials and methods. The theoretical part of the study was carried out using a classification method based on pattern recognition theory. A test image with an oil spill on the surface of the Black Sea was compared with the same image presented in BMP format with different color depth encodings, as well as with the same image presented in JPEG format. Raster images were processed using a specialized software application. Simulation was carried out in the MathCAD environment.Results. The developed approach was tested experimentally by processing 200 images. The conducted visual analysis confirmed that the image, where the given boundary allows the area of oil spill to be clearly distinguishes, is the closest to the test image.Conclusion. According to the results of the experiment with different formats of raster image files, the conclusion is made about the feasibility of using images obtained by visual control systems presented in JPEG format as initial data. Further research directions are outlined.

Keywords

• file format
• raster images
• oil product spill
• digital matrices