CAJ | 학술논문

变化检测方法作为遥感信息获取的主要技术手段之一，在灾害监测中已得到了一定应用。然而基于遥感图像灰度信息的变化检测方法，其阈值选取受到传感器和环境因素引起的“伪变化”信息影响。由于纹理特征具有移不变特性，可以利用地物空间结构信息开展变化检测，从而避免图像灰度“伪变化”信息的影响。该文在现有变化检测技术流程之上，通过纹理特征参数的分析，结合基于贝叶斯的图像阈值分割技术，建立了基于纹理特征的高分辨率遥感影像变化检测方法，并应用在汶川地震前后的福卫-2影像，与基于光谱影像的变化检测方法进行了分析对比。
변화검측방법작위요감신식획취적주요기술수단지일，재재해감측중이득도료일정응용。연이기우요감도상회도신식적변화검측방법，기역치선취수도전감기화배경인소인기적“위변화”신식영향。유우문리특정구유이불변특성，가이이용지물공간결구신식개전변화검측，종이피면도상회도“위변화”신식적영향。해문재현유변화검측기술류정지상，통과문리특정삼수적분석，결합기우패협사적도상역치분할기술，건립료기우문리특정적고분변솔요감영상변화검측방법，병응용재문천지진전후적복위-2영상，여기우광보영상적변화검측방법진행료분석대비。
Change detection using remote sensing images is one of the important methods on obtaining land feature information,and is widely used to monitoring the natural disaster.In the change detection methods based on image gray information,the threshold selection is affected by pseudo change information from the effect of sensors and environmental factors.Texture feature has a property of shift invariant,which can use the spatial structure in-formation in change detection and avoid the effect of pseudo change from image gray.An algorithm was presented that using the texture characteristic and image segmentation techniques.The algorithm was validated and analyzed using the sample data.The results show that the algorithm based on the texture characteristic can effectively extract the land features with a better precision.