CAJ | 학술논문

提出一种基于Sobel算子的球体IHS遥感影像融合算法，先对多光谱Mul影像做IHS变换并提取I分量，采用八方向模板Sobel边缘检测算子计算不同窗口参数（level）的全色Pan影像与I分量的局部区域清晰度值GP 和GI ，并根据局部区域的清晰度的大小确定不同区域的融合影像的I分量的新值I′。采用本文方法1（level＝12）和方法2（level＝24），通过实验并与其他融合方法比较，本文方法在提高影像分辨率的同时也保留了更多光谱信息和边缘细节，效果较好。
제출일충기우Sobel산자적구체IHS요감영상융합산법，선대다광보Mul영상주IHS변환병제취I분량，채용팔방향모판Sobel변연검측산자계산불동창구삼수（level）적전색Pan영상여I분량적국부구역청석도치GP 화GI ，병근거국부구역적청석도적대소학정불동구역적융합영상적I분량적신치I′。채용본문방법1（level＝12）화방법2（level＝24），통과실험병여기타융합방법비교，본문방법재제고영상분변솔적동시야보류료경다광보신식화변연세절，효과교호。
A sphere IHS remote sensing image fusion method based on Sobel operator is presented.First,the multispectral image Mul is transformed by IHS transformer and extracts I Component.After that,for different window parameter values (level)of the new remote sensing images fusion algorithm,eight direction template So-bel edge detection operators are used to calculate the local regional definition of the panchromatic Pan and the local regional definition of the I Component.According to the size of the local regional definition values,the new value of I Component of different parts of images is determined.When compared Method 1 (level=12)and Method 2 (level=24)with other fusion methods by experiments,it shows that Method 1 and 2 can effectively enhance spatial resolution while retraining more spectral information and edge details,the performance of the proposed methods is much better.