The Survey Of Segmentation And Central Location Of Typhoon Cloud Department

Typhoon is one of the disaster weather system that seriously influences the production and life of people. Typhoon brings the very great threaten and disasters to industry and agriculture production , traffic and transportation and the safe of life and property of people in cities of near southeast of sea in our country, the number of injured person is the highest in the ten great universe disasters, our country is one of the stricken serious country that is subjected to typhoon. Accurate forecast of typhoon is a big problem about the safety of our nation and people. The location of typhoon center is important to analy and forecast typhoon. The automatic location of typhoon center can improve the accuracy of location, and avoid the factor of person in the artificial location.This article takes the FY2 weather satellite cloud image as the research object. Because it is premiss to location of typhonic center that accurately segment out typhonic cloud department from various cloud departments. So this article firstly does segmentation of satellite image to remove non-typhoon cloud department. In this process, firstly, binary original color image to do the follow-up processing conveniently; then, using the opening of mathematical morphology and the feature of area remove the smaller non-typhoon cloud department; But at this time still exist the bigger non-typhoon cloud department in some cloud images. Because of this reason, this article puts forward to make use of the feature that typhoon cloud wall looks like the circle, namely eccentricity, in order to remove the non-typhoon cloud department that using the feature of area can not remove. Get the dense cloud wall region of typhoon finally.In the automatic location of typhoon center, for the sake of the improving the adaptability of algorithm and guaranteeing the accuracy of location, this article puts forward the method of subtration image, namely remove cloud wall from typhoon cloud department, in this processing, make use of the result of segmentation before location. So solved the problem that the phenomenon that spiral band joins cloud wall makes two different skeleton join together; and then, using the result of segmentation and original image extracted spiral band of typhoon; afterward, do the operation of fill holds before skeleton in order to avoid loop that is a blight to location; at last, fit helix to find out the position of typhonic center. Experiment proves the algorithms of segmentation and center location that put forward are effective.The research of the article can play a better role to forecast of typhoon, and has the actual applied value.