基于直觉模糊集的阔叶树苗叶片边缘检测

针对田间或苗圃植物背景的复杂性,为准确从采集的样本图片中分割出树苗叶片,提出了一种基于直觉模糊集的阔叶树苗叶边缘检测算法。首先采用3×3模板分别对RGB颜色空间中的R、G、B灰度图进行x方向、y方向、45°以及135°方向模糊聚类,然后采用最大类间方差法提取模糊聚类图像的阈值,最后根据阈值检测出阔叶苗叶边缘。对经典的基于微分算子的边缘检测法与该研究提出的边缘检测算法进行了分析比较,结果证明该研究提出的算法能较好地检测出阔叶树苗叶边缘,特别对于重叠区域叶片也能检测出边缘。

Edge detection of hardwood seedlings leaves based on intuitionistic fuzzy set

In this paper, a novel edge detection method based on intuitionistic fuzzy set(IFS)for hardwood seedlings leaves is proposed to accurately segment the leaves from images in complex background. Firstly, 3×3 templates of the x-direction, y direction, the direction of 45° and 135° were utilized to cluster the R, G, and B gray image respectively. Secondly, the maximum variance method was proposed to extract the fuzzy clustering image threshold. Finally, image edge was detected according to the threshold, where a clear boundary was obtained. Proposed method performed better than classical edge detection methods of differential operator. Experiments for kinds of hardwood seedlings were carried out, and the results indicated that the proposed method was effectiveness to detect the leaf edge, especially for the overlap region.