基于颜色特征的桔子树干图像分割方法

Image Segmentation Methods for Orange Trunk Based on Color Features

根据桔子树干颜色的特点,提出了一种图像分割方法。首先,利用颜色特征定位ROI;然后,计算该区域的颜色特征向量提取桔子树干;最后,对不连续的桔子树干区域利用数学形态学方法进行自动修补。实验结果表明,该方法能够有效地提取出桔子树干,并确定其质心和面积,算法的平均识别率达到了86.93%。

An image segmentation algorithm was proposed according to the characteristics of the Orange -tree trunk Firstly, the ROI was located by color features, Secondly, the Orange - tree trunk were vectors which were computed in ROI. Finally, some unconnected regions of the Orange extracted using the color feature -tree trunk were automatically repaired by Mathematical Morphology methods. The experimental results showed that this algorithm could effectively ex- tract the Orange - tree trunk and computed its centroid and area. The average accurate rate of image segmentation for the Orange-tree trunk was 86.93%.