Enhancement of night time video using dark channel prior ip accelerator

Abstract

Enhancement of night-time video using Dark Channel Prior IP accelerator is proposed in this thesis. Nighttime video processing is difficult due to low brightness, low contrast and high noise in the video. The above problems may affect the accuracy and may results in failures of object detection or object classification at night time. Dark Channel Prior (DCP) filter is used to improve the visibility, brightness, and contrast of the night-time input video. Processing speed is a challenging task on the real-time application of the DCP algorithm for night time video enhancement. Hence, the DCP algorithm is implemented on FPGA (ALVEO Board) to increase the speed of video processing. In order to demonstrate the quality of the proposed method, a practical environment with a Vitis software tool and Alveo board was also set up to evaluate the performance of hardware. Such extensive experimental results indicate that the proposed algorithm and hardware structure are effective, feasible and straightforward to apply to practice.