PRCA-based NSST image fusion method

Abstract

The present invention discloses a PRCA-based NSST image fusion method. The method comprises the following steps of S1, respectively subjecting to-be-fused images to multi-scale decomposition by using a non-down-sampling redundancy lifting non-separable wavelet (RLNSW), completing the directional localization and then obtaining the low-pass sub-band components and the band-pass sub-band components of the to-be-fused images; S2, fusing the low-pass sub-band components of the to-be-fused images based on the first fusion rule of the robust principal component analysis (RPCA), and fusing the band-pass sub-band components of the to-be-fused images based on the second fusion rule of the robust principal component analysis (RPCA); S3, after the fusion process of the step S2, subjecting the low-pass sub-band components and the band-pass sub-band components of the to-be-fused images to inverse NSST conversion, obtaining a fused image and outputting the image. The method is simple in implementation, good in image fusion effect, short in consumed time and good in practicability.

Description

technical field [0001] The invention relates to the technical field of multi-source image processing, in particular to an RPCA-based NSST image fusion method. Background technique [0002] With the rapid development of sensor technology, a single-mode sensor has developed into a sensor network or system composed of different imaging mechanisms, different working wavelength ranges, different working environments, and different working functional modules. The information obtained by the entire system presents diversity. volume increased sharply. The traditional single-mode information processing method is no longer applicable to the new problems brought by the multi-sensor combination. Data fusion technology is to combine all kinds of complex information obtained by multiple sensors into a new data set, so as to make full use of increasingly complex multi-source data for further analysis, processing and decision-making. Multi-source Image Fusion (Multi-source Image Fusion) i...

AI Technical Summary

Problems solved by technology

[0004] In image fusion processing, the fusion rules also have a great influence on the fusion effect. At present, the fusion rules are mainly divided into three categories: pixel-level, window-based, and region-based. Among them, the improved gradient projection non-negative space matrix analysis method (NMF) The fusion of low-pass sub-band components has a good effect, but for the band-pass sub-band components, the fusion effect is not good; for the fusion of infrared and visible light images, robust principal component analysis (RPCA) can obtain the sparse matrix of the original image , the background information of the infrared image is represented by the low-frequency component, and the target subject is represented by the sparse component, so the fusion effect of infrared and visible light images combined with RPCA is significantly improved

Some practitioners proposed an infrared-visible image fusion method based on RPCA and NSCT, which can improve the fusion effect, but cannot improve the fusion speed

[0005] For image fusion, there are mainly two types of fusion methods. The first type is the fusion method based on the space domain, such as linear weighting and neural network. The method of linear weighting is simple but the fusion effect is poor, and the method of neural network is good and universal. However, there are a large number of parameters to be determined, which are difficult to set accurately, and the fusion takes a lot of time due to the large amount of data and a large number of iterative operations; the second type is the fusion method based on the transform domain, such as the wavelet transform fusion method, Ridgelet transform fusion method, curvelet transform fusion method, contourlet transform fusion method and shearlet transform fusion method, etc. Among them, the non-subsampling shearlet transform fusion method has the most superior performance, but the fine detail of NSP multi-scale decomposition The capture ability is not strong, and the data volume of the bandpass subband components obtained by decomposition is large, which is not conducive to the rapid fusion of subsequent fusion rules

To sum up, the existing above-mentioned image fusion methods either have poor fusion effects or take too long, which seriously restricts the practical use of image fusion technology.

Images