A Hardware Friendly Haze Removal Method and Its Implementation

The images captured in hazy weather are always affected by haze, hence image dehazing has become a hot field in image processing. However, the existing dehazing algorithms are hard to be implemented on hardware to speed up because of high computational complexity. In this paper, we propose a novel hardware friendly image dehazing method to improve the image quality and save the execution time simultaneously. First, it selects airlight from the local patch with the current densest haze to obtain estimated airlight in real time. Then, it adopts the TDM (Time Division Multiplexing) strategy when searching the optimal transmission to save nearly 86.5% of hardware resources. Finally, it uses the fixed-point guided filter to improve the precision of the final refined transmission. The experimental results indicate it outperforms other dehazing algorithms in terms of SSIM and CIEDE2000. The hardware architecture for our proposed method trades off dehazing accuracy against throughput, which can obtain high-quality restored images with a high-throughput of 166. 67Mpixels/s when processing 80frames/s FULL-HD 1080p at 500MHz frequency.