CAJ | 학술논문

文章提出了一个图像区域分割算法。首先利用图像全局灰度分布特性聚类实现初始分割，然后进行区域标记和合并。新的合并方法建立在分析区域之间的几何和灰度关系的基础上，克服了现有的区域合并方法的局限性。算法应用于一些合成图像、自然景物和遥感图像，实验结果是令人满意的。
문장제출료일개도상구역분할산법。수선이용도상전국회도분포특성취류실현초시분할，연후진행구역표기화합병。신적합병방법건립재분석구역지간적궤하화회도관계적기출상，극복료현유적구역합병방법적국한성。산법응용우일사합성도상、자연경물화요감도상，실험결과시령인만의적。
A new region segmentation algorithm is presented in this paper. First,an image is initially segmented with a clustering way using the global intensity distribution property. Second,the regions are labeled and merged. The new merge method is based on analyzing the geometry and intensity relationship in order to deal with some drawbacks of the present methods. This method was applied to some synthetic images,nature scene images and remote sensing images,and the results are satisfying.