Research On Fusion Method Of Domestic High Spatial Resolution Satellite Images And Aerial Images

The remote sensing platform is equipped with different types of sensors,different sensors acquire images in different ways,and acquire multi-source image data with different characteristics.So how to effectively use these data to extract feature information? The study found that multi-source remote sensing data fusion technology has become an effective method,which can comprehensively use various image data to fully extract feature information.Similarly,the imaging methods of aerial photographs and satellite images are different,and the characteristics of the image data are different.The aerial photographs have higher spatial resolution than satellite images but lack spectral information.Satellite images acquire images from multiple channels.The images have rich spectral information.However,when information is extracted,researchers generally hope that the image has the characteristics of high resolution,rich spectral information,and rich spatial details.However,aerial photographs and satellite images have some of the characteristics.Therefore,we combine aerial photographs and satellite images.After extracting feature information,the obtained fusion image has the characteristics of complementary information: high resolution,rich spectral information,and rich spatial details.The fused image improves the accuracy of feature extraction in remote sensing images.It is known that the fusion of aerial photographs and high-resolution satellite images has very good application prospects and research value.(1)The system comprehensively analyzes the current development of remote sensing satellite and satellite image fusion algorithms,and the advantages and disadvantages of these fusion algorithms.In addition,it summarizes and analyzes the current status of fusion of remote sensing satellite images and aerial photographs,and discusses the problems existing in the fusion algorithms.(2)Four traditional remote sensing image fusion methods are discussed,namely IHS(Intensity-Hue-Saturation)fusion method,wavelet transform fusion method,PCA(Principal Component Analysis)transform fusion method,and Gram-Schmidt(GS)spectral sharpening fusion method.This article introduces the principles and fusion methods of the above four remote sensing image fusion methods,and performs fusion simulation experiments,and analyzes the advantages and disadvantages of each method.(3)Aiming at the problem of how to select the band in Unmanned Aerial Vehicle(UAV)photographs to replace the intensity component I during IHS fusion,this paper proposes a method of selecting the optimal band,which improves the IHS transform.The fusion algorithm experiment is programmed in Matlab environment,and the fusion image is evaluated from five aspects: spectral distortion,spatial distortion,spectral mapping feet,fusion quality evaluation,and fusion image correlation.The experimental results show that the proposed method optimizes the shortcomings of spectral loss and distortion caused by the fusion of the original IHS transform fusion method and the PCA transform fusion method to a certain extent.(4)In order to obtain the optimal fusion image based on UAV photographs and satellite images,an algorithm is proposed to effectively fuse drone aerial photos and high-resolution multispectral satellite images.Firstly,calculate the segmentation parameter Ve of the pulse-coupled neural network(PCNN)based on the grayscale information of the satellite image.Secondly,use the correlation between the UAV photographs' bands to reduce the UAV photographs to a single-band UAV photograph.The multi-spectral image is segmented by PCNN and the segmentation layer is edged.Finally,the injection weights of the multi-spectral image and the single-band UAV photograph are calculated to reconstruct a new multi-spectral image with rich spectral and spatial information.The proposed algorithm was evaluated using various high-resolution images,including Gao Fen-2(GF-2),GaoFen-1(GF-1)and UAV photographs.By comparing the algorithm in this paper with several classical fusion methods,research shows that the method is effective in fusing remote sensing images and UAV photographs.(5)Aiming at the problem of high resolution of UAV photographs and clear texture information,but insufficient spectral information,which is not conducive to the interpretation of drone aerial photos in later applications,a new fusion method based on PCA transformation is proposed(IPCA).Firstly,analyze the characteristics of the UAV photographs through the histogram of each band of the UAV photographs,thereby constructing a single-band UAV photograph-containing rich texture information and spectral information,and then replace it with a reconstructed single-band image.The first principal component in the PCA transformation process.Finally,the inverse PCA transformation is performed to finally obtain a fusion image with sub-meter spatial resolution and high spectral resolution.Experimental tests show that this method can effectively reduce spectral distortion and maintain texture details.