Image Dehazing Technology in Coal Mines Based on LDCNN-CVR
This paper proposed an efficient image dehazing method based on LDCNN-CVR(Light-DehazeNet Convolutional Neural Network-Color Visibility Restoration),aiming at the problems of the low contrast and blurred details of images captured in mine video surveillance.Firstly,it used the transformed atmospheric scattering model through a lightweight CNN frame-work to jointly estimate the transmission map and atmospheric light for single-image dehazing and noise reduction.Then,it proposed a color visibility restoration method to reduce the color distor-tion in the dehazed images.Finally,we tested the proposed method and compared the experiment data to verify that the dehazing strategy of this paper outperformed other algorithms in subjective and objective evaluations,which was suitable for dehazing enhancement of images to obtain high-quality reconstructed images in coal mines.
single image dehazingimage reconstructionconvolutional neural networkcolor visibility restoration
万方数据
维普
