Real-Time Implementation of Object Detection Algorithm for Underwater Environment

An underwater environment where lighting and visibility are generally not controllable causes the underwater image of objects to be distorted as well as blurred most of the times. Therefore, the valuable information from such an image cannot be fully extracted for further processing. To solve the problems mentioned above and to improve the quality of underwater images, pre-processing of images is attracting tremendous interest [7]. This paper presents a new method based on bilateral transform along with Contrast equalization i.e., adaptive histogram equalization technique to preserve the edges of objects more effectively and efficiently. Bilateral transform is used for de-noising based on a soft threshold. SURF is used to extract the features. Qualitative analysis indicates that the proposed method exhibits better performance in enhancing the quality and retaining the edges of an image by cancelling neighbouring unwanted information [7].