An Adaptive Fusion Method of SAR Image and Multispectral Image

Abstract

The invention relates to a self-adaptation SAR image and multispectral image fusion method. According to the self-adaptation SAR image and multispectral image fusion method, firstly, BEMD is carried out on a source image, the problem that the correlation of local features of image fusion based on wavelet transformation is not powerful is solved, and the problem that a wavelet basis function is difficult to select with a traditional wavelet method is avoided; secondly, high and low frequency selection is carried out according to intrinsic mode function (BIMF) components obtained through T detection, low-frequency information is fused based on the area self-adaptation weighting information criterion, the area matching degree is calculated based on the high-frequency component, coefficients with the matching degrees in the different threshold value ranges are fused, threshold values are selected based on the particle swarm optimization, and the defects caused when the area matching threshold values are determined according to the experience are avoided; lastly, the fusion result is obtained in the manner that BEMD inverse transformation is carried out on the fusion components. Thus, the combination of BEMD and the particle swarm optimization can greatly improve the quality of the fused image, and has importance significance and using value on subsequent processing and image displaying of an application system.

Description

technical field [0001] The invention belongs to the field of data information fusion, and in particular relates to an adaptive SAR image and multispectral image fusion method, which can be applied to various military or civilian SAR image and multispectral image fusion systems. Background technique [0002] Due to the great difference between SAR image and multispectral image in terms of imaging mechanism and spectral characteristics, previous fusion methods often have problems such as low fusion contrast, loss of important information, and serious spectral distortion, which makes it difficult to achieve satisfactory fusion of SAR image and multispectral image. fusion effect. Wavelet-based image fusion cannot design more effective fusion rules based on the correlation characteristics between transformed coefficients, resulting in obvious noise in some areas after fusion, and cannot preserve the details of multispectral images well. . In addition, wavelet-based and wavelet-...

AI Technical Summary

Problems solved by technology

[0002] Due to the great difference between SAR image and multispectral image in terms of imaging mechanism and spectral characteristics, previous fusion methods often have problems such as low fusion contrast, loss of important information, and serious spectral distortion, which makes it difficult to achieve satisfactory fusion of SAR image and multispectral image. fusion effect

Wavelet-based image fusion cannot design more effective fusion rules based on the correlation characteristics between transformed coefficients, resulting in obvious noise in some areas after fusion, and cannot preserve the details of multispectral images well.

In addition, wavelet-based and wavelet-like decompositions rely on pre-defined filters or wavelet functions, and image fusion based on wavelet transform is to fuse the local features of each pixel or small area, which will lead to loss of local feature correlation. strong feature

Images