Particle detection on low contrast image of large aperture optics
Ding Wendong; Xu D(徐德); Zhang ZT(张正涛); Zhang DP(张大朋)
2016-05
会议名称2016 Chinese Control and Decision Conference (CCDC)
会议日期2016-5
会议地点yinchuan
摘要It's a challenge to identify actual deposited particles from the acquired optics image mixed with large noises. Due to the long working distance constraint for online inspecting large aperture optics, the vision system has great depth of view. It results in that the acquired optics image is blended with stain contamination on the back side of the inspected optics surface. In this paper, we propose a particle detection method for the low contrast image of large aperture optics. The method consists of three steps including particle candidate detection, image alignment, and particle determination. The particle candidate detection algorithm combines the gradient information to detect small particles. The actual particles are determined by the subtraction of the reference image from the inspected image with the topological information. For the detected particle, this paper also gives a classification method to identify dust and defect. Experiment shows that the actual particle on the low contract image of large aperture optics can be effectively detected and classified with the proposed method.
关键词Particle Inspection Image Alignment Reference Image Subtraction Classification.
文献类型会议论文
条目标识符http://ir.ia.ac.cn/handle/173211/19821
专题精密感知与控制研究中心_精密感知与控制
作者单位Institute of automation, chinese academy of sciences
推荐引用方式
GB/T 7714
Ding Wendong,Xu D,Zhang ZT,et al. Particle detection on low contrast image of large aperture optics[C],2016.
条目包含的文件 下载所有文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
07531929.pdf(569KB)会议论文 开放获取CC BY-NC-SA浏览 下载
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[Ding Wendong]的文章
[Xu D(徐德)]的文章
[Zhang ZT(张正涛)]的文章
百度学术
百度学术中相似的文章
[Ding Wendong]的文章
[Xu D(徐德)]的文章
[Zhang ZT(张正涛)]的文章
必应学术
必应学术中相似的文章
[Ding Wendong]的文章
[Xu D(徐德)]的文章
[Zhang ZT(张正涛)]的文章
相关权益政策
暂无数据
收藏/分享
文件名: 07531929.pdf
格式: Adobe PDF
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。