Snow Category Extraction of NOAA/AVHRR Images by using Three Dimensional Histogram HARAMOTO Yoshiaki, IZUHARA Muneto, KAZI A Kalpoma Graduate School of Information Science Tohoku University Sendai, Japan Jun-ichi Kudoh Center for Northeast Asian Studies Tohoku University Sendai, Japan Abstract— It is important to judge the presence of the snow for prevention of the snow damage to which crops and livestock suffer. However, because of the few meteorological observation point, enough information is not obtained in the plateau region like Mongolia. In this research, the snow area is classified only by the NOAA/AVHRR image assuming application in Mongolia, using three dimensional histogram, and using neither the specialist's judgment nor the grand truth data. The validity of this method is examined compared with snow information that the specialist classified, as the preliminary research in the Tohoku region in Japan. In this method, snow data inside the NOAA images are collected by visual observation, and the snow database is constructed. However, because the image with the snow was few, the data base of an enough size was not able to be constructed. Then, the form of a snow category is extracted using the advantage in which three dimensional histogram was able to be analyzed interactively, and a snow area was classified. As for verification, Miss-classification rate is that the difference between classified snow area by this method and by specialist’s method is divided by the specialist's classification. Miss- classification rate became about 50% as a result. Furthermore, Miss-classification area found that it concentrated on the boundary of the snow. In the boundary, snow is very vague with a resolution of NOAA/AVHRR. When 1 pixel is permitted from the boundary, Miss-classification rate became about 25%. And when three pixels permitted, it became about 11%. From the above, snow could be classified in the considerable precision by using neither specialist's judgment nor ground truth data. It is applicable in the area with few meteorological observation points like Mongolia, by using this method. The possibility of prevention of damage, such as early precaution of the snow damage, was able to be found. (Abstract) Keywords - snow coverage ; Remote Sensing ; Mongolia ; NOAA/AVHRR I. INTRODUCTION Recently in the Mongolian area, livestock have suffered seriously by the snow damage called zud. In this research, by using satellite image snow in earth surface is detected by aiming at early warning and reduction of this snow damage. Kazama et al [1]-[4] studied an excellent research over the existence of the snow coverage of earth surface, the depth of snow coverage, the seasonal variation by using the satellite image and the ground observation data which are called AMeDAS (Automated Meteorological Data Acquisition System). However, in a highland area like Mongolia, since there are few weather observation points, sufficient information is not acquired. Moreover, there are few images with the snow and the snow data cannot be collected enough training data. Therefore, simple collation between unknown data and training data can not be performed. In this research, we develop the new snow detection method which does not use ground observation data but use only the satellite images by using three dimensional histogram Algorithm developed by Kudoh et al [5][6]. In the Tohoku region in Japan, the classification is tried for comparison of the results with Kazama et al, as a preliminary research. II. THREE DIMENSIONAL HISTOGRAM The method using the three dimensional histogram is one of the supervised classifications, a category only with a little domain can also be treated. Moreover, there is an advantage that interactive processing is possible since it excels in visualization of classification process, and it is a method intelligible also for a non-specialist's user. This research detects snow area solving the problem where snow data cannot be collected easily by using this interactive analysis. III. DATA This research uses the NOAA-14 data from February 1995 to March 2001. NOAA is a weather satellite series being operated by National Oceanic and Atmospheric Administration (NOAA). NOAA is a solar synchronous, almost true circle semi-polar orbit satellite. And the orbital inclination is about 98 degrees. The altitude of 870km and 833km, and about 100 minute/cycles, and regress 9 to 10 days. NOAA carries Advanced Very High Resolution Radiometer (AVHRR). There are five channels (Ch) with sensitivity from a visible to the far- infrared wavelength in AVHRR. The resolution of all channels is 1.1 km directly under the satellite and observation width is about 2800km. The radiation intensity detected by AVHRR is recorded with the discrete integer value (0-1023; 10bit count). Ch1 and Ch2 correspond to albedo (%), and Ch3-Ch5 corresponds to brightness temperature (K). 0-7803-8743-0/04/$20.00 (C) 2004 IEEE