Study on the seasonal stability of remote sensing image fusion in nearshore areas

WANG Yunxia; MA Anqing; YAN Xia

doi:10.13984/j.cnki.cn37-1141.2025.03.006

Abstract：In order to investigate the seasonal stability of remote sensing image fusion models in the nearshore region,the Landsat panchromatic(PAN)data and MODIS multispectral(MS)data from four seasons were fused in five ways:nearest neighbour diffusion(NNDiffuse),ratio transform(Brovey),colour transform(HSV),phase recovery(Gram-Schmidt)and principal component analysis(PCA).The quality of the fused images was evaluated in terms of qualitative,quantitative and applica-bility,and the seasonal stability of the remote sensing fusion model was analyzed using variance(S2).The results showed that in terms of qualitative evaluation,NNDiffuse,Brovey and HSV had high sea-sonal stability and good fusion effect.In terms of quantitative evaluation,nine quantitative evaluation indexes were selected.Using normalization to calculate the integrated evaluation index,the HSV fusion model S2 was 2.4× 10-3,which had high stability and could significantly improve the remote sensing image quality in all seasons in the nearshore region.In terms of the inversion accuracy of chlorophyll concentration,the S2 of HSV fusion model was 4.9 × 10-3,which had high stability among the four seasons and was correlated well with the measured chlorophyll concentration(R2＞0.619).Therefore,the HSV fusion model with high seasonal stability,good fusion quality and high accuracy of chlorophyll inversion was the optimal model for nearshore remote sensing image fusion.