基于LSD的具有特定特征的三维物体轮廓提取算法

王宇,王硕,谭民

系统科学与数学 ›› 2016, Vol. 36 ›› Issue (9) : 1410-1416.

PDF(552 KB)
PDF(552 KB)
系统科学与数学 ›› 2016, Vol. 36 ›› Issue (9) : 1410-1416. DOI: 10.12341/jssms12896
论文

基于LSD的具有特定特征的三维物体轮廓提取算法

    王宇,王硕,谭民
作者信息 +

THE CONTOUR EXTRACTION ALGORITHM BASED ON LSD DETECTOR FOR A 3-D OBJECT WITH SPECIFIC FEATURE

    WANG Yu ,WANG Shuo, TAN Min
Author information +
文章历史 +

摘要

提出一种基于LSD的具有特定特征的三维物体轮廓提取算法, 该算法主要用于检测结构环境中的特定目标, 提取其轮廓目的在于计算出目标的位姿信息, 便于机器人作业抓取. 针对具有特定特征的三维物体, 算法首先对物体进行基于LSD的外轮廓提取, 其次再对原始图像进行基于改进的Prewitt梯度算子的边缘提取, 并删除经边缘提取后外轮廓特征信息, 再进行基于LSD的物体内轮廓提取, 从而实现物体的轮廓提取. 基于LSD的轮廓提取算法包括提取梯度图像, LSD线段检测, 直线段的分类和轮廓的拟合四个步骤. 在直线段分类中, 提出了利用极坐标表示线段特征, 利用极径和极角对线段进行有效分类的方法. 为了进一步证明三维轮廓提取算法的有效性, 以长方体为例, 详细地阐述了该算法的具体步骤.

Abstract

In this paper, we propose a contour extraction algorithm based on LSD detector for a 3-D object with specific feature, the algorithm is used to detect the designated target in the structured environment. The contour of the designated target is applied to compute its pose, and let the robot grasp it. First, the outer contour extraction based on LSD detector is achieved, then the image edge is detected using the improved Prewitt gradient operator, and the outer contour feature is deleted. Thus the inner contour is extracted, and the total contour extraction is realized. The contour extraction algorithm based on LSD detector consists of gradient image extraction, LSD detector, line segment classification and contour fitting. Segment classification based on polar coordinates is proposed for the line segment classification. In order to show the effectiveness of three-dimensional contour extraction algorithm, taking a cuboid for an example, we give the description for the specific steps of the proposed algorithm.

关键词

3-D轮廓提取 / 特征提取 / LSD线段检测.

引用本文

导出引用
王宇 , 王硕 , 谭民. 基于LSD的具有特定特征的三维物体轮廓提取算法. 系统科学与数学, 2016, 36(9): 1410-1416. https://doi.org/10.12341/jssms12896
WANG Yu , WANG Shuo , TAN Min. THE CONTOUR EXTRACTION ALGORITHM BASED ON LSD DETECTOR FOR A 3-D OBJECT WITH SPECIFIC FEATURE. Journal of Systems Science and Mathematical Sciences, 2016, 36(9): 1410-1416 https://doi.org/10.12341/jssms12896
中图分类号: 93C95   
PDF(552 KB)

226

Accesses

0

Citation

Detail

段落导航
相关文章

/