Abstract
The dark channel prior (DCP)-based single image removal algorithm achieved excellent performance. However, due to the high complexity of the algorithm, it is difficult to satisfy the demands of real-time processing. In this article, we present a Graphics Processing Unit (GPU) accelerated parallel computing method for the real-time processing of high-definition video haze removal. First, based on the memory access pattern, we propose a simple but effective filter method called transposed filter combined with the fast local minimum filter algorithm and integral image algorithm. The proposed method successfully accelerates the parallel minimum filter algorithm and the parallel mean filter algorithm. Meanwhile, we adopt the inter-frame atmospheric light constraint to suppress the flicker noise in the video haze removal and simplify the estimation of atmospheric light. Experimental results show that our implementation can process the 1080p video sequence with 167 frames per second. Compared with single thread Central Processing Units (CPU) implementation, the speedup is up to 226× with asynchronous stream processing and qualified for the real-time high definition video haze removal.
Highlights
Light attenuation induced by haze causes an image or video captured by camera to undergo subjective and objective information loss
Based on the resources provided by RESIDE, we evaluate nine representative stateof-the-art algorithms: Dark-Channel Prior (DCP) [3], Fast Visibility Restoration (FVR) [44], Boundary Constrained Context Regularization (BCCR) [45], Artifact Suppression via Gradient Residual Minimization (GRM) [46], Color Attenuation Prior (CAP) [47], Non-Local
We presented a real-time Graphics Processing Unit (GPU) implementation for a haze removal algorithm based on dark channel prior (DCP)
Summary
Light attenuation induced by haze causes an image or video captured by camera to undergo subjective and objective information loss. The performance of computer vision algorithms (e.g., feature detection, target recognition, and video surveillance) inevitably suffers from the degraded low-contrast scene radiance. Removing haze can significantly increase the visibility of the image and correct the color bias caused by atmospheric light. Haze removal algorithms are attracting much attention and being widely studied in computer vision applications. Researchers proposed a series of algorithms aimed to remove the haze from noise-free images. Tan [1] suggested that haze-free images have higher contrasts than those with haze, so the haze can be removed by increasing the contrast of the image
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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 call
