Research On Modling Of Calculatmg Cloud-top-height And Cloud-vector Based On Phototopography

Cloud-top-height and cloud-vector retrieved by satellite data are two kinds of important data in synoptic meteorology and climatology. However, there are still a lot of problems existed in the calculation of Cloud-top-height and cloud-vector, which limits the application of Cloud-top-height and cloud-vector. Now, how to improve cloud-top-height and cloud-vector has become a crucial question to be solved. More accuracy cloud-top-height and cloud-vector could highly improve the quality of weather analysis and mathematical forecast, which means a lot to scientific research.The calculation method of cloud-top-height and cloud-vector by using stereogram based on meteorological satellite is a new and highly precise calculation method based on geometrical parameters. Currently, cloud-top-height is solved by using re-projection firstly, then by using maximum wind speed or average wind speed as correction term. The accuracy of this method is limited by the accuracy of re-projection, and wind speed and cloud-top-height is solved, separately.Based on collinearity equation, this paper introduced the parameter of cloud movement speed, and established the coupling model of cloud-top-height and cloud-vector and the exactly calculation method of cloud-top-height and cloud-vector. Then, this paper offered a bidirectional matching method based on correlation coefficient method, which is based on image intensification and Cloud(Snow) detection according to single channel threshold, owing to the low resolution and the cloud shadows of Medium-Resolution image. We also calculated the cloud-top-height and cloud-vector in cloudy areas, snow areas and cloudy-snow areas by using stereogram constructed by multi-time and multi-angel (AN,AF and AA) satellite images and coupling model of cloud-top-height and cloud-vector. The conclusions are as follows.(1)Cloud(Snow) detection based on single channel threshold can effectively remove the effect of cloud movement and cloud shadows, and can improve matching speed and matching success rate; the information content of medium resolution satellite image in cloud(snow) areas are added and the rate of coverage in matching homonymy cloud (snow) points is improved by using image enhancement method. Therefore, the matching success rate between AN, AF and AA image can achieve100%, and the maximum error is less than2pixels.(2)NASA insists that MISR Cloud-top-height products without wind speed is more stable and more precise; the result of the first generation of MISR product is generally higher, because this product takes maximum wind speed as correction term. However, the result of the new generation of MISR product is getting better because it takes average wind speed as correction term.(3) Used coupling model of cloud-top-height and cloud-vector to calculate the ground control points. The theoretical error of cloud-top-height and cloud-vector is approximately300m, respectively; The mean square error(MSE) in terms of theory of cloud-top-height and cloud-moved-distance is below300meters through adjustment methond of indirect observations; The D-value is almost300m after comparing MISR product without wind with our result; The D-value between second generation of MISR product and our result, is also below300m, which once again proves that the error of cloud-top-height we calculated is300m; Mass data detection proves that coupling model of cloud-top-height and cloud-vector is stable and can be used to distinguish cloud and snow.(4)The error is approximately1m/s after calculating the wind speed of ground points; the wind direction of continuous cloud points are almost the same, and their wind speed are really close, which consists with the ground observation.(5)Accomplished total cloud cover product with the resolution of275m, by using cloud-top-height to calculate total cloud cover. This kind of product can not only better perform the details of Cloudy areas and non-cloudy areas, but also distinguish Snow areas with Cloud areas, which can makes up for the high cloud cover caused by the erroneous judgement of snow or cloud.The creative points of this study include:(1)Established the coupling model of cloud-top-height and cloud-vector based on collinearity equation and put forward precise calculation method for calculating cloud-top-height and cloud-vector. This method offers a new way to quantitative calculate cloud parameters in Medium-Resolution image, because this method is based on geometric principle rather than physical characteristics of cloud. (2) Analyzed existed problems in matching homonymy points in multi-angle image when we are respectively using correlation coefficient method, least square method or relaxation method. Owing to the low resolution and the cloud shadows of Medium-Resolution image, we therefore provided a bidirectional matching method based on correlation coefficient method, which is based on image intensification and Cloud(Snow) detection according to single channel threshold.