Research On Spatio-Temporal Fusion Method Of Multi-Source Remote Sensing Data By Introducing Regional Growth Theory

High-quality remote sensing data provides strong data support for various dynamic changes of the earth.However,due to the limitations of satellite technology and cost control,it is difficult to obtain high-quality remote sensing data.Spatio-temporal fusion of multi-source remote sensing data is the main method to obtain high-quality remote sensing data.The research of spatiotemporal fusion methods based on multi-source remote sensing data to generate high-quality remote sensing data is of great significance for the study of various dynamic changes of the earth.This paper proposes a spatiotemporal fusion method that introduces regional growth theory,selects some typical areas of Shijiazhuang and Yueyang as the research area,selects Landsat data and MODIS data as the research data,adopts a spatiotemporal fusion model that introduces regional growth theory,and obtains through experiments Fusion results and evaluation of fusion accuracy.The main work and conclusions are as follows:(1)Using the core idea of the regional growth model,according to the first law of geography the principle of adjacent similarity,make full use of the reflectance information in adjacent similar pixels to reduce the intra-class heterogeneity between similar pixels and central pixels though improve the existing remote sensing data spatio-temporal fusion model ESTARFM(improved ESTARFM).(2)Taking the observation image as the reference object,the results are analyzed from both qualitative and quantitative perspectives.Qualitatively,the improved ESTARFM has a good visual effect;quantitatively,in experimental area A,except for the near red band,the correlation coefficient of the improved ESTARFM in the other bands is higher than that of ESTARFM to varying degrees,and the root mean square error is lower than ESTARFM;peak signal-to-noise ratio is better than ESTARFM model;structural similarity is above 0.9.In experimental area B,the correlation coefficients of the improved ESTARFM in the blue,green,near-red,and shortwave infrared bands are better than those of ESTARFM;in the blue,green,and red bands,the root mean square error of the improved ESTARFM is less than ESTARFM;the peak value of the improved ESTARFM The signal-to-noise ratio is better than ESTARFM in the blue,green,red,and short-wave infrared bands;the structural similarity is all above 0.9.(3)The fusion efficiency of the improved ESTARFM and the original ESTARFM model in two different regions was compared and analyzed.The results showed that in experimental area A,the fusion time of the improved ESTARFM was 1:47:19.279 seconds,while the fusion of the ESTARFM model took time at 5h,20 m and 24.315 seconds;in experimental area B,the fusion cost of the improved ESTARFM was 20 minutes and 12.576 seconds,while the fusion time of the ESTARFM model was 23 minutes and 31.375 seconds;the time consuming of the improved ESTARFM fusion was much less than that of the ESTARFM model.The spatio-temporal fusion method that introduces the regional growth theory considers the impact of inter-class heterogeneity and also considers the problem of intra-class heterogeneity.In two different experimental areas,the improved ESTARFM model has better effects than ESTARFM,namely The band correlation coefficient is improved,the peak signal-to-noise ratio and the root mean square error are reduced,and the structural similarity is above 0.9,which greatly improves the fusion efficiency and provides a certain reference and reference for large-scale spatio-temporal fusion research.