Underwater image enhancement using improved Particle Swarm Optimization

The underwater images are used for monitoring the underwater pipelines and aquaculture. However, obtaining a clear image is a difficult task due to light scattering and absorption issues. Due to these issues, a very low contrast, blur, and color distortion underwater image is obtained. To overcome these issues, enhancement algorithms have been designed. In the literature, conventional algorithms are used for enhancement purposes are histogram equalization, stretching, and power-law expression. However, these algorithms over-enhance the images if proper parameters are not taken under consideration. Therefore, swarm-based optimization algorithms have been used to determine the optimal values of these parameters. In this paper, we have designed an improved particle swarm optimization-based enhancement algorithm. In the proposed method, initially, pre-processing of the underwater image is done and its channels are extracted. After that, based on the mean values of the channels, the channels are classified into three types, namely, superior, intermediate, and inferior. Based on the superior channel, gain factors are determined and enhance the intermediate and inferior channels. Next, the power-law expression is applied to the inferior and intermediate channels to enhances its gamma value. The gamma value in the proposed algorithm is determined using an improved particle swarm optimization (PSO) algorithm. In the improved PSO algorithm, the initial population of the PSO algorithm is defined using a chaotic map algorithm therefore PSO algorithm does not fall into the local optimal solution. In the last, contrast-limited adaptive histogram equalization (CLAHE) for enhancement. The experimental results are performed for the standard dataset images and simulated in MATLAB. Further, qualitative and quantitative analysis is done for the proposed algorithm. The results show that the proposed algorithm is superior as compared to the existing algorithms