Underwater Image Processing to Detect Polymetallic Nodule Using Template Matching
Prabhakaran K, Ramesh R, Nidhi Varshney, Rajesh S, K. Gopakumar, Gidugu Ananada Ramadass, M A Atman · Global Oceans 2020: Singapore – U.S. Gulf Coast · 2020
Underwater image processing is extensively different from visual image processing on land, dueto scattering of light, absorption and refraction. It is very difficult to get clear and recognizable underwater images, because of back-scattering, noise, blurriness, turbidity, lower brightness& contrast, etc. This paper deals with exploring techniques for detection of Polymetallic Nodules abundance lying on the sea bed. In order to recognize the polymetallic nodules on the sea bed and to quantify, we propose three stages of processing on optical image captured by AUV by applying simple and effective techniques such as histogram, histogram equalization and CLAHE through template matching approach with the creation of multi-scale templates from available underwater optical images with different lightings. Through this approach, the matching objects are detected with less error and retrieved more appropriate results and quantified the number of nodules available in the given area. As a testing process, hundreds of multi-scale templates are created from different source images in different lightings and tested with source and non-source images which furnished better results by setting the suitable threshold value. The processing was done using existing hardware and no extra electronics devices and sensors were used.