Video Defogging with a Sample-Based Defogging Approach

Increased atmospheric particulate matter has resulted in increased fog and haze, which has reduced visibility and presented a significant challenge for computer vision applications. Photos taken under these conditions typically have low contrast and poor visibility, making object detection, tracking, and recognition more difficult. As a result, improving the clarity and contrast of photos and videos—also referred to as de-fogging—is necessary. Real-time de-fogging is important, and the minimum preserving sub-sampling-based technique works well for foggy videos because of its effectiveness and suitability for a wide range of image types. This technique integrates scene change detection to preserve temporal and spatial coherence while controlling computational costs. Experimental results demonstrate that this video de-fogging approach achieves excellent spatial and temporal coherence.