Remote Sensing Image Dehazing Using Heterogeneous Atmospheric Light Prior

Remote sensing images (RSIs) captured in haze weather will suffer from serious quality degradation with color distortion and contrast reduction, which creates numerous challenges for the utilization of RSIs. To address these issues, this paper proposes a novel removal algorithm, named HALP, visible RSIs based on heterogeneous atmospheric light prior side window filter. HALP is comprised two key components. Firstly, given large imaging space RSIs, treated as globally non-uniform distribution instead global constant. Therefore, simple effective method estimation presented, utilizes brightest pixel each local image patch region. Secondly, filter-based transmission algorithm proposed, can effectively suppress block effect map caused by minimum filter used dark channel algorithm. Experiments both real-world synthetic remote demonstrate effectiveness HALP. In terms no-reference full-reference assessments, yields excellent results, outperforming existing state-of-the-art algorithms, including physics-based neural network-based methods. The visual comparison dehazed results also shows that restore degraded uneven haze, producing clear rich details natural colors.