Atmospheric Light Estimation using Particle Swarm Optimization for Dehazing

For the past decade, many researchers have been working towards the improvement in the visibility of single hazy images, using the haze image model. According to the haze image model, the hazy-free image is restored by estimating the atmospheric light and transmission from a hazy image. The objective of this proposed work is to improve the perceptibility by decreasing the density of haze in the hazy images. The research work was carried to estimate the optimal value of atmospheric light by tuning the weights using a bioinspired technique called Particle Swarm Optimization (PSO) based on the objective of minimizing the fog density. We have selected a fitness function or objective function which incorporates all statistical features to differentiate a clear image from the hazy image. The results are validated with the state-of-the-art, by measuring fog density of the restored image using Fog Aware Density Evaluator (FADE). Also, the results are validated by measuring the Peak signal to noise ratio (PSNR) and structural similarity index (SSI) using ground truth images from Foggy Road image database (FRIDA). This research work demonstrates better results qualitatively and quantitatively.

@article{N2019,
  title     = {Atmospheric Light Estimation using Particle Swarm Optimization for Dehazing},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {10},
  number    = {11},
  year      = {2019},
  publisher = {The Science and Information Organization},
  author    = {Padmini T N and Shankar. T},
  doi       = {10.14569/IJACSA.2019.0101175},
  url       = {https://doi.org/10.14569/IJACSA.2019.0101175}
}