Ecological and environmental problems have become increasingly prominent in recent years. Environmental problems represented by haze have become a topic that affects the harmonious ecology of human beings. The trend of this topic is on the rise. People's perception of the environment after the impact of haze has also changed. A real-time grasp of the dynamic public environment perception of emotions is often an important basis for environmental management departments to effectively solve environmental problems through public opinion. This article focuses on the problem of the public perception of emotional changes, which is caused by fog and hazy weather, proposes an environmental emotion perception model, using Weibo comment data about fog and haze as environmental perception data, and analyzes the impact of fog and haze on the public in four seasonal time dimensions. The post-environment perception of emotion changes: the results show that in spring, the public's environmental perception of emotions is mainly negative emotions at the beginning of the season; in summer, positive emotions become dominant emotions; in autumn, the public's environmental perception of emotions is dominated by negative emotions that increase substantially; and in winter, the dominant environmental perception of emotions of the public is still negative. This theory provides support for research on social emotions and public opinion behavior.
Keywords: convolutional neural network; deep learning; environmental perception of emotion; weighted networks.