Monitoring systems are rapidly becoming a frequently used technology in the smart town to increase the quality of human life in digital living, where rover finding is one of the critical factors for people- centric smart megacity operations such as wellbeing, security, traffic guiding, and unmanned vehicles. Bad weather, snow and rain have a great impact on traffic enforcement and pedestrians streaming. For tracking pedestrians, our system employs two steps as main idea techniques: picture decomposition and dictionary learning. As beginning, a guided filter in conjunction with rain/snow detection is used. Then, obtain visual information from the high-frequency section by designing a hierarchy of three layers structure. During the top layer, the overarching lexicon is learned. In the second layer, the result is exposed to a second guided filtering in conjunction with rain/snow detection. A topmost layer estimates the sensitivity to change across color

Image enhancement, guided filtering, Structural SimilarityIndex (SSIM), Peak Signal to Noise Ratio (PSNR).