In this work, we Dark Channel Priorpropose a parallel method that applies the dark channel priorDark Channel Prior to implementation in a digital signal processor to eliminate fog in a scene image. Therefore, our research is particularly relevant because it aims to improve haze removal algorithms, especially for video surveillance and self-driving vehicles, where clear vision is crucial. Dark Channel PriorDark Channel Prior, a widely used method, estimates the medium transmission function of an image by identifying the minimum intensity value for each pixel. A haze-free image typically contains one RGB channel with intensities close to zero and an intensity proportional to the haze in the original image. However, processing time remains a significant challenge for digital devices. To address this, we propose a parallelization approach to optimize computational resources, implemented on a digital processor programmed in C. Performance was evaluated using metrics such as mean absolute error, peak signal-to-noise ratio, and processing time on a dedicated dataset. Results show that although the Python algorithm achieved marginally better numerical performance, the proposed parallelization method reduced processing time by 17.34%, demonstrating its potential for enhanced computational efficiency.