ABSTRACT
Poor weather conditions like fog and haze are major concern in many computer and vision applications. Images captured in poor weather are degraded in its various features like edge, contrast, saturation and color. These images cannot be used directly in applications like object detection, tracking and surveillance (security). It requires pre-processing so that degraded features can be restored to improve quality of foggy image. In this chapter, white balancing approach is used as a pre-processor to extract illuminant color of foggy image. Guided filter is used in intermediate stages to preserve features like edges, textures and fine details. In the final stage, partially overlapped sub-block histogram equalization has been used to enhance contrast of restored image. Qualitative and quantitative analyses have been done on both natural and synthetic data sets with state-of-the-art method.
