-
公开(公告)号：KR1020080022257A
公开(公告)日：2008-03-11
申请号：KR1020060085444
申请日：2006-09-06
Applicant: 아주대학교산학협력단
CPC classification number: G06K9/6211
Abstract: A method for classifying leafs by using venation feature points is provided to extract the venation feature points and classify and store the leafs according to the venation having the extracted feature points, thereby enabling a user to search a plant exactly having a venation including similar feature points. A method for classifying leafs by using venation feature points comprises the following steps of: extracting a sample image for a venation of a leaf through a extracting unit(S401); extracting a BP(Branching Point) and an EP(Ending Point) as feature points by applying a CSS(Curvature Scale Space Corner Detection) algorithm to the sample image(S405); classifying the BP and the EP through an analyzing unit(S407); calculating density of the extracted feature points through non-parametric estimation method by using a Parzen window, and confirming whether the feature points are distributed in a linear form or distributed around one point(S409); confirming vertical distribution of the feature points, determining the feature point as a parallel venation if the feature points are distributed around upper and lower one point, and determining the feature point as a venation if the feature points are distributed in the linear form(S410); if parallel venation, confirming vertical and horizontal distribution of the BP(S420) and determining whether density of the BP is upwardly inclined and lower density is distributed in a linear form(S425); and if the lower density of the BP is present in the linear form, classifying the feature points as a second parallel venation, and classifying the feature points as a first parallel venation if the density of the BP is upwardly distributed and the lower density is distributed around one point(S427,S428).
Abstract translation: 提供了一种通过使用静脉特征点对叶片进行分类的方法，以提取脉搏特征点，并根据具有提取的特征点的脉搏对叶片进行分类和存储，从而使用户能够精确地搜索具有包括相似特征点的脉搏的植物 。 通过使用静脉特征点对叶片进行分类的方法包括以下步骤：通过提取单元提取用于叶片的静脉的样本图像（S401）; 通过对样本图像应用CSS（Curvature Scale Space Corner Detection），提取BP（分支点）和EP（终点）作为特征点（S405）; 通过分析单元对BP和EP进行分类（S407）; 通过使用Parzen窗口通过非参数估计方法计算所提取的特征点的密度，并确定特征点是以直线形式分布还是分布在一个点附近（S409）; 确定特征点的垂直分布，如果特征点分布在上下一点的情况下，将特征点确定为平行脉冲，并且如果特征点以线性形式分布，则将特征点确定为静脉（S410） ; 如果平行脉动，确认BP的垂直和水平分布（S420），并确定BP的密度是否向上倾斜，并且较低的密度以线性形式分布（S425）; 并且如果BP的较低密度以直线形式存在，则将特征点分类为第二平行脉冲，并且如果BP的密度向上分布且较低密度分布，则将特征点分类为第一平行脉冲 约一点（S427，S428）。
-
公开(公告)号：KR100821671B1
公开(公告)日：2008-04-14
申请号：KR1020060085444
申请日：2006-09-06
Applicant: 아주대학교산학협력단
CPC classification number: G06K9/6211
Abstract: 본 발명은 각종 식물의 잎맥 이미지에 곡률기반의 코너 검출기법을 이용하여 잎맥의 특징점을 추출한 후, 추출된 잎맥의 특징점의 밀도를 파젠 윈도우를 통해 비모수 밀도 추정방법으로 계산하여 잎을 분류 및 저장하여, 해당 잎맥을 갖는 식물에 대한 정확한 검색을 용이하게 하는 잎맥의 특징점을 이용한 잎 분류방법에 관한 것이다.
-
Search Results
2
records
(took 0.017 seconds)
