Research On Ground Object Information Recovery Algorithm For Thin Cloud Remote Sensing Images Based On Image Fusion

Satellite remote sensing sensors are easily affected by the weather when obtaining ground images.Ground object information of remote sensing images covered with thin clouds is obscure or missing,which seriously affects the interpretation and analysis of the image.In the current situation,a very effective method of ground objects information recovery is an important way to enhance the availability of cloud contaminated remote sensing images.This paper researches on ground object information recovery algorithms for thin cloud remote sensing images based on image fusion by using thin cloud remote sensing images acquired by Landsat and HJ-1 satellite.The main contents and achievements are as follows:1.Describes the purpose and significance of information recovery algorithm for ground objects in thin cloud images.Introduces the basic parameters of each band of Landsat series and HJ-1 A/B/C satellites sensors.And the present situation of the research on information recovery for ground objects at home and abroad are introduced.2.The physical characteristics of clouds and the cloud features in the remote sensing images are introduced.The basic theories and methods of remote sensing image pre-processing are elaborated,in order to prepare for further research.3.An information recovery algorithm for ground objects in thin cloud images based on PCNN fusion and T-SVR is proposed.Because there is a certain difference in ground object information of the remote sensing images in the same area at different phases,thin cloud remote sensing image is firstly the change detection.Then the dual-tree complex wavelet transform is used to decompose region without change,Then different PCNN fusion methods are separately used to enhance the high frequency subbands and suppress the low frequency subbands.The low frequncy subbands of changed region decomposed by dual-tree complex wavelet transform are predicted by T-SVR.Finally,the ground object information recovery image is reconstructed.4.An information recovery algorithm for ground objects in thin cloud images is proposed by fusing guide filter and transfer learning.Firstly,multi-resolution decomposition of thin cloud target images and cloud-free guidance images is performed by using multi-directional nonsubsampled dual-tree complex wavelet transform.Then the decomposed low frequency subbands are processed by using support vector guided filter and transfer learning respectively.The decomposed high frequency subbands are enhanced by using modified Laine enhancement function.The low frequency subbands output by guided filter and those predicted by transfer learning model are fused by the method of selection and weighting based on regional energy.Finally,the enhanced high frequency subbands and the fused low frequency subbands are reconstructed by using inverse multi-directional nonsubsampled dual-tree complex wavelet transform to obtain the ground object information recovery images.