296
Asian Journal of Engineering, Sciences and Technology - Volume 10, Issue 2 2020
By Syed Safdar Hussain, Hamid Saeed Khan
De-focusing of lenses and bokeh are two causes of blurry images in an underwater setting. The noise created by the water environment obscures the object's features in the photograph. Integration of three different image processing techniques has been made to recognise object in blur image with the least amount of false feature matching. Preprocessing to reduce image blurriness and noise, Euclidean form detection to reduce the risk of false matching, and feature extraction for object detection are all included. A supervised machine learning strategy for Euclidean form detection has been suggested to prevent classifiers from making false matches. Features are extracted from those detected Euclidean shapes, which not only decrease the chance of false matching but also Features are retrieved from the recognised Euclidean shapes, which raises the confidence that a geometrically shaped object is present in the image while reducing the possibility of false matches.