摘要
分析了BP神经网络结构和基于大气传输系统的图像退化模型,通过将雾天图像频谱作为大气传输逆系统的输入,将同一场景下的晴天图像频谱作为其输出,建立对应该大气逆系统的BP模型,并据此对雾天图像进行清晰化处理,消除不良天气对图像质量的影响。
The back propagation neural network structure and image degradation model based on atmospheric modulation transfer system are analyzed in this paper. We get a BP model of the atmospheric modulation transfer inverse system by the relationship of the image frequency spectrum extracted from an image in fog and the one in good weather condition in the same background. By determining the weights of network, the bad weather effects can be eliminated and the fog image can be sharpened.
出处
《电视技术》
北大核心
2012年第19期44-46,共3页
Video Engineering
基金
四川省教育厅项目(08ZC029)
关键词
大气调制传输逆系统
神经网络
最大熵
atmospheric modulation transfer inverse system
neural network
maximum entropy
