目的 为了实现草坪杂草管理的精准化施药，针对自然环境中杂草与草坪颜色相近导致杂草难以分割的问题，提出一种改进模糊C均值(Fuzzy C-means, FCM)聚类的分割算法。方法 利用超绿算子提取感兴趣区域，融合HSV空间的多通道信息进行图像预处理，扩大杂草与草坪的特征差异。使用区域面积约束滤波范围，去除预处理图像中的草坪背景噪声，降低中值滤波造成的目标区域灰度级损失。提出一种各向灰度分布差异(Difference of gray distribution, DGD)检测算子，在聚类过程中引入像素周围不同方向的灰度分布差异特征实现草坪杂草分割。结果 与传统FCM、FCM-S2、FCMNLS以及RSFCM算法相比，本文算法对大多数噪声区域抑制效果较好，可以实现较为理想的杂草分割效果。本文算法能有效分割草坪杂草，平均分割准确率达到91.45%，比FCM、FCM-S2、FCMNLS和RSFCM算法分别提高16.35%、4.12%、6.80%和8.06%。结论 本文算法可有效地分割自然环境中的草坪杂草，为草坪杂草精准化施药提供了条件，具有实际应用价值。
Objective In order to realize the precise application of herbicides for lawn weed management, an improved fuzzy C-means (FCM) clustering segmentation algorithm was proposed to solve the problem that it is difficult to segment weeds due to the similar color between weeds and lawns in natural environment.Method The region of interest was extracted by extra-green operator, and the multi-channel information in HSV space was incorporated for image preprocessing to expand the feature difference between weeds and lawns. The filtering range was constrained using region area to remove the lawn background noise in the preprocessed image and reduce the gray level loss in the target region caused by median filtering. An anisotropic detection operator of difference of gray distribution(DGD) was proposed. In the clustering process, the gray distribution difference characteristics in different directions around pixels were introduced to realize lawn weed segmentation.Result Compared with the traditional FCM, FCM-S2, FCMNLS and RSFCM algorithms, the algorithm in this paper (DGDFCM) had better suppression effect on most noise areas and could achieve ideal weed segmentation effect. The DGDFCM algorithm could effectively segment lawn weeds, and the average segmentation accuracy was 91.45%, which was 16.35%, 4.12%, 6.80% and 8.06% higher than FCM, FCM-S2, FCMNLS and RSFCM algorithms, respectively.Conclusion The DGDFCM algorithm can effectively segment lawn weeds in natural environment, provides the condition of precise application of herbicides for lawn weeds, and has a practical application value.