Abstract
Image haze removal is critical for autonomous driving. However, it is a challenging task for the existing image dehazing algorithms to eliminate the block effect completely and handle objects similar to light (such as snowy objects and white buildings). To address this problem, we propose a novel single-image dehazing method based on superpixels and Markov random field. We obtain the transmission map in the superpixel domain to eliminate the block/halo effect and introduce Markov random field to revise the transmission map in the superpixel domain. The key idea is that the sparsely distributed, incorrectly estimated transmittances can be corrected by properly characterizing the spatial dependencies between the incorrectly estimated superpixels and the neighbouring well-estimated superpixels. The experimental results demonstrate that the proposed method outperforms state-of-the-art image dehazing methods.
Highlights
Outdoor images taken in hazy weather have low contrast and visibility
Model-based image recovery methods come from physically valid algorithms that remove haze by modelling the optical transmission of imaging in scattering media and the prior information to remove the backscattered light in front of the scene and to compensate for the light attenuation of the scene
This paper proposes a new approach for the estimation of the transmittance using a superpixel and Markov random field (MRF) model
Summary
Outdoor images taken in hazy weather have low contrast and visibility. The low image quality will greatly affect human perception of colours in the image and humans’ object recognition and matching abilities. INITIAL TRANSMISSION MAP ESTIMATION BASED ON the DARK CHANNEL PRIOR AND SUPERPIXELS The optical model commonly used to describe the image observed in a scattering medium is. The proposed method successfully preserves the edge details while reducing the artefacts (Fig. 2(c)) To address this problem, we segment an input image into numerous superpixels by simple linear iterative clustering (SLIC) [16], and we define (x) as the superpixel to which pixel x belongs. Yang et al [17] and Noh et al [18] calculated the transmittance by using the minimum value of superpixels It is impossible for the boundaries of superpixels to always adhere to the structural edges exactly, which means that a normal minimal operation may yield a block effect, which is less serious but still evident than that from He’s method (Fig. 3). The number of superpixels is set to 3000 in our experiments
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
