Research On Image Fusion Based On Compressed Sensing

Image fusion is the synthesis of image information of the same scene acquired by multiple sensors or the same sensor in different modes to obtain more accurate,comprehensive,and reliable images.Multi-focus image fusion refers to the incorporation of clear objects from multiple different focus areas into aoptimal image.Remote sensing image fusion in this paper refers to the fusion between panchromatic images and multispectral images,which can preserve the multispectrality while improving the spatial resolution of multispectral images.The explosion of information has made people more and more demanding of information processing technology,and the requirements for image quality have become higher and higher.As the signal carrying information grows wider and wider,the signal processing based on the Nyquist sampling theorem requires increasingly higher sampling rates and processing speeds,and the processing of wideband signals becomes more difficult.Compressive sensing theory breaks through the limitations of the traditional Nyquist sampling theorem.The theory proposes that a set of specific waveforms are used to perceive the signal,the signal is projected to a given waveform,a set of compressed data is perceived,and finally the solution optimization problem is used.From the compressed observation data,rebuild the original data.Under this theory,information fusion between source images can be performed at the level of observations,which greatly reduces the amount of data fusion,data storage,and transmission costs.This paper studies the image fusion based on compressed sensing.In this paper,three parts of the compressive sensing: sparse representation,compressed observation,and algorithm reconstruction are discussed in detail.On the basis of image fusion based on compressive sensing,the effect of compressive sensing algorithm on the fusion image is not ideal.Wavelet decomposition is introduced and an image fusion algorithm based on wavelet decomposition-compressive sensing is proposed.Experiments are conducted on multi-focus and remote sensing images.First,the image is decomposed by wavelet,and the high-frequency components are combined based on compressive sensing.After averaging the low-frequency components or taking a panchromatic images,the fusion image is obtained after inverse wavelet transform.After the fusion experiments of multi-focus and remote sensing images were performed,the fusion results were analyzed and it was verified that the improved algorithm had a better fusion result.Image fusion based on compressed sensing represents an important branch of image fusion,and it has very important strategic significance in remote sensing,medicine,and national defense.