植物叶裂特征自动提取研究  

The Study on Automatic Feature Extraction of Leaf Lobes

作　　者：郑小东[1] 张晓煜[1] 薄树奎[1] 

机构地区：[1]郑州航空工业管理学院计算机科学与应用系,郑州450015

出　　处：《中国农学通报》2012年第27期152-156,共5页Chinese Agricultural Science Bulletin

基　　金：河南省科技攻关项目"基于图像处理的植物叶特征描述"(102102110149);国家自然科学基金项目"遥感影像单类信息提取中类别划分研究"(41001235);航空科学基金项目"自动目标识别中尺度问题研究"(2011ZC55005)

摘　　要：为了实现叶裂特征自动提取,构造了植物叶片凸包距离函数,通过分析该函数,判断是否存在叶裂,如果存在叶裂,进一步判别叶裂类型,获取叶裂数量与深度数据。利用树叶数据库中的图像对方案进行了测试,结果显示:叶裂存在性参数对叶片是否具有叶裂区分明显,叶裂位置检测、叶裂数目判定准确,叶裂类型判别参数体现了不同叶裂类型的差异,叶裂深度参数体现了叶裂的深度差异。测试表明方案切实可行,可用于植物的种类自动识别和长势监控系统。A function named convex hull distance was constructed in order to obtain automatic feature extraction of leaf lobes. After analysis of the function, it was judged if there were leaf lobes or not. If the result was yes, the type of leaf lobes would be identified and the quantity and depth of leaf lobes would be acquired according to the function step by step. The scheme had been tested using images in tree leaf database. The results showed that the parameter of leaf lobes existence made a clear distinction between leafs with leaf lobes or not; the locations of leaf lobes on leaf contour were found precisely and the judgments on leaf lobes number were right; the parameter of leaf lobes type showed obvious difference between different leaf lobes type; the parameter of leaf lobes depth showed difference of leaf lobes in depth. The tests showed the scheme was feasible. It could be applied to systems of automatic plant identification and plant growth monitoring.

关 键 词：叶裂 特征提取 凸包 离散余弦变换 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]