Design of an FPGA-Based High-Quality Real-Time Autonomous Dehazing System

Abstract

Image dehazing, as a common solution to weather-related degradation, holds great promise for photography, computer vision, and remote sensing applications. Diverse approaches have been proposed throughout decades of development, and deep-learning-based methods are currently predominant. Despite excellent performance, such computationally intensive methods as these recent advances amount to overkill, because image dehazing is solely a preprocessing step. In this paper, we utilize an autonomous image dehazing algorithm to analyze a non-deep dehazing approach. After that, we present a corresponding FPGA design for high-quality real-time vision systems. We also conduct extensive experiments to verify the efficacy of the proposed design across different facets. Finally, we introduce a method for synthesizing cloudy images (loosely referred to as hazy images) to facilitate future aerial surveillance research.

Adaptive wavelet transform defogging scheme for real-time video restoration with field programmable gate array implementation

“…12 that the results of Liu et al 12 and Upadhyay and Sarawadekar 13 are darker, such as the wall of the classroom image. Similarly, the results of Tan et al 14 and Lee et al 16 are also suffering from over-enhancement. Our results have more balanced colors and sharp object details overall, whereas the performance on FPGA is similar to that on PC.…”

“…Tan et al 14 also have color distortion in the sky area, and the brightness of some object areas is dark after defogging. When the fog is not very large, Lee et al 16 will cause excessive fog removal, resulting in oversaturated colors. Our defogging approach can get adequate satisfactory defogging results in most cases, even in sky scenes or places where the depth changes abruptly.…”

“…8 Image processing based on FPGA is expected to meet the requirement of real time video defogging. [9][10][11][12][13][14][15][16] Liang et al designed a real-time hardware accelerator for single image defogging on FPGA. 10 It takes around 13.74 ms at 100 MHz when processing a 720 × 576 image, and the real-time performance is poor.…”

“…Sahu et al 15 proposed a dehazing method by cascading two models utilizing a novel parameter-adaptive dual-channel modified simplified pulse coupled neural network. Lee et al 16 presented an FPGA-based autonomous dehazing system that could handle real-time DCI 4K images/videos. Although the method can reduce color distortion and halo effects occurred on defogging processing, it consumes a large amount of logic and storage resources.…”

Automatic Deployment of Convolutional Neural Networks on FPGA for Spaceborne Remote Sensing Application

“…Meanwhile, spaceborne remote sensing platforms not only perform CNN-based image processing, but also perform image preprocessing, such as radiation correction and image dehazing, to improve the performance of CNN-based image processing [24]. Some FPGA-based studies showed that image preprocessing requires a lot of hardware resources to implement [25,26]. For example, Qi et al [25] implemented onboard image preprocessing for optical images on FPGAs.…”