Implementation of Histogram Based Image Fusion Technique for Underwater Image Enhancement in Reconfigurable Platform

摘要

Objectives: This paper proposes a novel histogram based image fusion method to refine the visibility of underwater images which are suffered from colour degradation and poor illumination due to ocean depth. Methods/Statistical Analysis: To improve the enhancement of underwater images, R channel is separately processed as red component attenuates heavily compared to blue and green component and is combined with histogram equalized original image. Then various weight maps, suitable for underwater images, were obtained from the single image and are processed before fusion is done. Findings: Automated Underwater Vehicles (AUV) is used for marine navigation, pipeline tracking, video mosaicking etc. Automation of video captured has to be used for various applications where image pre processing is necessary. Hardware implementation helps the automation process to be used in real time scenario. In this work Field Programmable Gate Array implementation (FPGA) of a novel algorithm for the underwater image enhancement is carried out. New weight maps suitable for underwater images were used for fusion as compared to existing methods and were compared with the existing methods. Altiumnano board was used for FPGA implementation. Application/Improvements: Degraded images need to be enhanced before processing it further applications. This work is mainly focussed on underwater images and can be used in AUVs and Remotely Operated Vehicles (ROV).