基于彩色图像分割的孔隙度提取

Extraction the Canopy Porosity Based on Color Image Segmentation

为了研究彩色图像分割技术测定树木叶面积指数(LAI)的可行性,利用彩色图像聚类方法获取冠层图像的孔隙率。提出一种方法克服FCM聚类中心和聚类数对初始条件的敏感性问题。首先使用彩色直方图提供聚类算法初始条件;然后将结果应用到FCM算法中分割冠层图片;最后将实验结果与can-eye软件测量结果比较。结果表明,它们之间的相关系数为0.93,达到显著或极显著水平,说明该研究方法为获取森林LAI提供了有效途径。

In order to study the technical feasibility of using color image segmentation to measure trees＇ leaf area index（ LAI）,the canopy porosity based on color image segmentation is extracted. A new method is proposed to overcome the sensitivity of the initial conditions of FCM cluster centers and cluster numbers. First,the initial conditions of clustering algorithm are provided by using color histogram. Then,the results are applied to the FCM algorithm to segment the canopy image. Finally,the experimental results are compared with the results from the Can-eye software measurement.The experimental results show that the correlation coefficient is 0. 93,which achieves significant or extremely significant level. It is illuminated that the improved FCM algorithm is an effective way to obtain the porosity of the forest canopy and leaf area index（ LAI）.