摘要
图像预处理是零件特征提取与识别的基础,处理质量直接决定后期识别的效果。提出了基于灰色关联分析的图像分割新算法,该算法通过分析像素点序列与代表目标的参考序列的灰色关联度来进行区域分割。通过测量实验,证明其对于在较复杂背景图像中较模糊的目标边缘具有较好的检测效果。可完整的提取出目标区域,并得到连续封闭的目标边缘,为后续的零件目标识别打下了良好的基础。
Image preprocessing is fundamental to the feature extraction and recognition of spare parts. The quality of the preprocessed image determines the effectiveness of the subsequent recognition. In the paper,a new image segmentation algorithm base on grey incidence analysis is discussed and attempted,in which grey incidence of the current pixel series and the reference series representing the object is applied to the regional segmentation. Measurements and experimentations show that the algorithm exhibits good performance for images with considerably complex background and blurred edge. The object region can be extracted completely and a closed and connected border can be obtained,which has laid a good foundation for the subsequent object recognition of the spare parts.
出处
《组合机床与自动化加工技术》
北大核心
2014年第10期63-65,共3页
Modular Machine Tool & Automatic Manufacturing Technique
基金
河北省科技计划项目(13211815)
关键词
零件识别
图像分割
灰色关联分析
spare part recognition
image segmentation
grey incidence analysis
