Technology Research Of Vegetation Information Extraction Based On Fully Polarimetric SAR Image

As an important part of the earth's ecosystem,vegetation is closely related to human activities.The extraction of vegetation information is of great significance for monitoring environmental change,agricultural development and prediction of geological disasters.Especially in the mountain area,the vegetation information can be used as an important index to predict landslides and debris flows.The traditional method of acquiring vegetation information is very limited,mainly through two kinds of methods: manual measurement and optical remote sensing.Manual measurement is difficult to execute in some condition and large data is difficult to obtain.Optical remote sensing is easy to be affected by the weather.SAR technology has the characteristics of cloud,fog and rain,it can monitor the surface information all day.Because of the penetrability of microwave,it has a good distinction between vegetation and other objects,and it is easier to retrieve vegetation biomass information.This paper is mainly about the research on the extraction technology of vegetation biomass information of fully polarimetric SAR image based on target polarization decomposition method,mainly includes research,vegetation cover information extraction technology and information extraction of vegetation biomass.Specially as follows:(1)Cloude decomposition for scattered region plane scattering mechanism of lower scattering coefficients mistakenly decomposed into high entropy scattering mechanism.This is due to the low SNR which resulting in the polarization channel bulk scattering coefficient difference decreased.(2)In this paper,the method of polarization azimuth compensation is used to reduce the influence of topography on the decomposition of Yamaguchi.(3)This paper combined the H/ alpha-Wishart classification method with Otsu method,effectively improve the classification results.Compared with the traditional H/alpha-Wishart classification method,the improved method is able to retain more details..(4)This paper proposes a Yamaguchi-SVM classification method based on region,which improved classification results for complex terrain from 62.4% to 71.3%..(5)In this paper,a complete information extraction process of vegetation biomassbased on full polarimetric SAR imagery is implemented in two regions of Qionglai and Zhaojue.The mean square error of Qionglai study area is 0.662,the correlation coefficient is 0.893;the result of Zhaojue study area has 75% pixel values of the experimental points in the experimental range,the maximum offset to less than 1.6.