Using a hybrid of interval type-2 RFCMAC and bilateral filter for satellite image dehazing

Cheng Jian Lin, Chun Hui Lin, Shyh Hau Wang

研究成果: Article同行評審

3 引文 斯高帕斯(Scopus)


With increasing advancement of science and technology, remote sensing satellite imaging does not only monitor the Earth’ s surface environment instantly and accurately but also helps to prevent destruction from inevitable disasters. The changing weather, e.g., cloudiness or haze formed from atmospheric suspended particles, results in low contrast satellite images, and partial information about Earth’ s surface is lost. Therefore, this study proposes an effective dehazing method for one single satellite image, aiming to enhance the image contrast and filter out the region covered with haze, so as to reveal the lost information. First, the initial transmission map of the image is estimated using an Interval Type-2 Recurrent Fuzzy Cerebellar Model Articulation Controller (IT2RFCMAC) model. For the halo and color oversaturation resulted from the processing procedure, a bilateral filter and quadratic function nonlinear conversion are used in turn to refine the initial transmission map. For the estimation of atmospheric light, the first 1% brightest region is used as the color vector of atmospheric light. Finally, the refined transmission map and atmospheric light are used as the parameters for reconstructing the image. The experimental results show that the proposed satellite image dehazing method has good performance in the visibility detail and color contrast of the reconstructed image. In order to further validate the effectiveness of the proposed method, visual assessment and quantitative evaluation were implemented, respectively, and compared with the methods proposed by relevant scholars. The visual assessment and quantitative evaluation analysis demonstrated good results of the proposed approach.

期刊Electronics (Switzerland)
出版狀態Published - 2020 五月

All Science Journal Classification (ASJC) codes

  • 控制與系統工程
  • 訊號處理
  • 硬體和架構
  • 電腦網路與通信
  • 電氣與電子工程


深入研究「Using a hybrid of interval type-2 RFCMAC and bilateral filter for satellite image dehazing」主題。共同形成了獨特的指紋。