Multisource Remote Sensing Imagery Fusion Scheme Based On BEMD And The Extraction Of Bamboo Forest In Li River

Imaging mechanism of Synthetic aperture radar(SAR)is different from optical sensors.SAR is able to acquire shape and structure characteristics of physiognomy and work in all weather and all day condition.Optical panchromatic images have high spatial resolution so that it can clearly reflect the edge feature.Multi images have high spectral resolution,which is helpful to the visual interpretation.Fusion of SAR and optical images can fully complement the information of data as well as reduce the uncertainty of classification.In this paper,we study the fusion scheme of SAR and optical images based on bidimensional empirical mode decomposition(BEMD).The main contents are as follows:(1)Investigating the development history and background of image fusion method.The research focus on empirical mode decomposition(EMD)and bidimensional empirical mode decomposition(BEMD)in signal processing and application principle and application in remote sensing image fusion which lays the foundation for the follow-up work.(2)BEMD has many advantages in dealing with nonlinear and non-stationary signals and are with full self-adaptability.This paper design and implement many sets of remote sensing image fusion and comparative experiments based on BEMD,which covered:the contrast of HSV transform fusion,GS transform fusion,wavelet transform fusion and BEMD fusion by using SAR image and multispectral image;the contrast of curvelet transform,a-trous transform,DWT wavelet transform fusion and BEMD fusion by using SAR image and panchromatic image;the contrast of HSV transform fusion,GS transform fusion,wavelet transform fusion and BEMD fusion by using panchromatic image and multispectral image.Making full use of the limited data resources.And evaluated the fusion results by visual interpretation and computational quantitative indicators.This has reached the maximum of information fusion.The experimental results show that the fusion method in this paper has advantages in the integration of SAR and optical images,and fully complements the SAR and optical image features,the fusion effect is better than other fusion methods.(3)Lijiang River is the main eco-tourism resources in Guilin.Due to human activities in recent years has intensified and over exploitation of tourism resources in the Lijiang River,ecological balance of Lijiang River had destroyed seriously,and the bamboo forest has increased rapidly.In this paper,we extract bamboo from fusion images by Gray-level co-occurrence matrix(GLCM).In fifth chapter we fused SAR image,panchromatic and multispectral images based on BEMD and compared with the fusion of panchromatic image and multi spectral image and the fusion of panchromatic image with SAR image.We optimize the GLCM based on semivariable function,thus the extract accuracy of bamboo has been improved.