Research On Multi-source Image Fusion System Based On DSP

As one of the most common information in daily life,image quality is increasingly required.Although current sensors have good performance,they often only capture a specific piece of information.At this time,in order to obtain an image with complete and accurate information,multi-source image fusion technology is needed.At present,this technology has been applied well in all walks of life,among which medical and health services,military and national defense,and intelligent transportation are more common.In the past research related to image fusion,the focus of attention is mostly how to improve the quality of image fusion,rarely involving the efficiency of image fusion.Therefore,this paper makes a theoretical study to improve the efficiency of image fusion on the premise of ensuring the quality of image fusion.Firstly,the related theories of image fusion are introduced synthetically,in which the evaluation standard of image fusion is mainly studied,and the subjective evaluation and objective evaluation are explained concretely,and six quantitative indexes are selected as the objective evaluation standard of this simulation experiment.Secondly,DWT,NSCT and NSST in multi-scale transform are mainly introduced,their advantages and disadvantages are compared,and simulation experiments are carried out.The results show that the image fusion quality of DWT algorithm is the lowest,the image fusion quality of the NSCT algorithm and the image fusion quality of the NSST algorithm are both good.Moreover,the NSST algorithm also has a high computational efficiency.Thirdly,the sparse transformation,observation sampling and reconstruction of compressed sensing theory are studied in detail.Then an image fusion algorithm based on NSST-CS is proposed,and the fusion rules of different subbands are designed respectively by using the regional characteristics.The simulation results show that the running time of the algorithm is reduced by 50%and the image fusion efficiency is greatly improved due to the application of compression sensing theory.Finally,according to the characteristics of image fusion algorithm,the hardware circuit of the corresponding real-time image processing system based on DSP is designed.Figure[34]Table[5]Reference[60]...