基于多策略融合技术的水稻叶片边缘检测算法研究  

Rice leaf edge detection algorithm based on multi-strategy fusion technique

作　　者：何丹丹[1] 徐梅[2] 

机构地区：[1]东北农业大学工程学院,哈尔滨150030 [2]东北农业大学理学院,哈尔滨150030

出　　处：《东北农业大学学报》2012年第5期65-68,共4页Journal of Northeast Agricultural University

基　　金：东北农业大学科学技术项目(2011RCA01)

摘　　要：农作物边缘检测问题是基于图像处理技术的农作物检测技术的重要内容之一。在分析已有边缘检测方法的基础上,将图像区域最小外接矩形算法、中值滤波、Canny算子和闭运算有机结合,提出一种基于多策略融合技术的水稻叶片边缘检测算法,将该方法应用于单叶片、背景为白板的叶片、背景为土壤的叶片、重叠叶片等不同特征的图像,有效地获取连续、光滑的水稻叶片边缘,表明该算法具有高效、准确和鲁棒的特性。Crops edge detection is an important content of crop detection technique based on image processing. On the basis of analyzing the existing edge detection methods, this paper combined the minimum circumscribed rectangular algorithm, median filtering, canny edge detectors with closed operation, and then proposed a rice leaf edge detection algorithm based on multi-strategy fusion technique. This algorithm was applied in different images, which included single leaf, overlap leaves, and leaves on the background of whiteboard and soil. A continuous and smooth rice leaf edge was obtained. The results showed that this method is effective, accurate and robust.

关 键 词：水稻叶片边缘检测 中值滤波canny算子 数学形态学 

分 类 号：S24[农业科学—农业电气化与自动化] TP391.4[农业科学—农业工程]