中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
机构
采集方式
内容类型
发表日期
学科主题
筛选

浏览/检索结果: 共2条,第1-2条 帮助

条数/页: 排序方式:
Detection of Large Herbivores in UAV Images: A New Method for Small Target Recognition in Large-Scale Images 期刊论文  OAI收割
DIVERSITY-BASEL, 2022, 卷号: 14, 期号: 8, 页码: 22
作者:  
Ma, Jiarong;  Hu, Zhuowei;  Shao, Quanqin;  Wang, Yongcai;  Zhou, Yanqiong
  |  收藏  |  浏览/下载:28/0  |  提交时间:2022/09/26
A new algorithm of image segmentation for overlapping grain image (EI CONFERENCE) 会议论文  OAI收割
ICO20: Optical Information Processing, August 21, 2005 - August 26, 2005, Changchun, China
作者:  
Zhang X.;  Zhang X.;  Zhang X.
收藏  |  浏览/下载:10/0  |  提交时间:2013/03/25
Image segmentation is primary issue in image processing  at the same time it is principal problem in low level vision in computer vision field. It is the key technology to process image analysis  image comprehend and image depict successfully. Aim at measurement of granularity size of nonmetal grain  a new algorithm of image segmentation and parameters calculation for overlapping grain image is studied. The hypostasis of this algorithm is present some new attributes of graph sequence from discrete attribute of graph  consequently achieve that pick up the geometrical characteristics from input graph  and new graph sequence which in favor of image segmentation is recombined. The conception that image edge denoted with "twin-point" is put forward  base on geometrical characters of point  image edge is transformed into serial edge  and on recombined serial image edge  based on direction vector definition of line and some additional restricted conditions  the segmentation twin-points are searched with  thus image segmentation is accomplished. Serial image edge is transformed into twin-point pattern  to realize calculation of area and granularity size of nonmetal grain. The inkling and uncertainty on selection of structure element which base on mathematical morphology are avoided in this algorithm  and image segmentation and parameters calculation are realized without changing grain's self statistical characters.