Precipitation estimates from microwave imagers have been the best to draw the global precipitation map. However these sensors are on low earth orbit (LEO), thus temporal sampling is a major drawback. A simple technique to interpolate rain rate observed from LEO microwave imagers or to fill the gaps which microwave data are not available is presented by using moving vectors from geo-IR data. The technique is evaluated using Radar AMeDAS data sets near Japan. The results show that the moving vector technique reduces the error by 0-30% compared to a simple linear interpolation.