【作者】 王敏；

【导师】 傅圣雪；

【作者基本信息】 中国海洋大学 ， 信号与信息处理， 2010， 硕士

【摘要】 目前对降雨事件的短时预测,获得了广泛关注,特别是对中尺度系统,当然也包括了雷暴情况。因为雷暴天气通常会伴随着天气在空间和时间范围内的迅速改变,并且会带来天气灾害,其影响包括从农民到飞行员等各种职业。所以对雷暴的发展、演变和运动情况的精确预报是至关重要的。本文在小波融合理论的基础上,首次在小波融合算法中考虑图像数据的物理意义,将可见光图像高的空间分辨率和丰富的纹理细节信息融合到红外图像中,提高了红外图像的空间分辨率,增加了红外图像数据的物理信息量。本文所提出的方法,充分考虑了红外和可见光图像本身数据的物理意义,用小波变换得到待融合数据的多分辨结构,在数据多分辨结构相应各级上采取不同的融合规则,得到融合数据的多分辨结构,利用小波逆变换重构融合数据。此方法得到的高分辨率红外图像不仅在视觉上得到了改善,而且丰富了红外通道数据的物理信息,取得了非常好的效果。结果表明：该融合方法简便易行,精度较高,为快速获取高分辨率红外图像提供了一条新途径。再对高分辨率的红外数据进行分析处理,预报出潜在对流初生(Convective initiation, CI)区域,使CI的预报时间更加提前。通过对我国京津地区一些时段的MTSAT卫星云图资料进行实验分析,预报CI。同时使用相同区域、同一时间段的雷达回波数据作为降水资料,对CI的预报结果进行检验,并根据检验结果对八分法系统参数进行调整,开发出一套适合中国地域的CI预报方法,最后取得了比较好的预报结果。更多还原

【Abstract】 One aspect that has received widespread attention is the short-term prediction of rainfall events, especially those that evolve on meso-time scales, which certainly include thunderstorm convection. Because thunderstorms are often accompanied by rapidly changing weather in spatial and temporal scales, and thunderstorms will bring weather hazards and phenomena that often adversely impact on professionals ranging from farmers to pilots. So there is a critical need to accurately predict their develop-ment, evolution, and movement.Based on the theory of wavelet fusion, this paper consideres the physical meaning of image data in wavelet fusion algorithm for the first time. This algorithm fuses the visible light images with high spatial resolution and rich texture detail to the infrared image, which can improve the saptial resolution of the infrared images, and increase the physical information. The method presented in this paper consider the physical meaning of the infrared and visible images. This paper uses multiwavelet to get the multi-resolution structure of the original data, and uses different fusion rules at every data multi-resolution structure level to get the multi-resolution stucture of the fusion data, which can be rebuild. This method not only get high-resolution infrared images, but also make the infrared channel data rich in physical information. The results show that:the fusion method is simple and has high precision, which is a new way of getting high-resolution infrared images.Then high-resolution infrared data were processed and analysed for the purpose of forecasting the potential CI region and a more earlier forecasting time. After fore-casting CI through the analysis of MTSAT image data of Beijing-Tianjin region, using radar data of the same time and the same area to test the CI forecasting result and adjusting the system parameters, a CI forecasting method was developed, which is suitable for China’s geographical and obtained good prediction results.更多还原