基于早期认知视觉的图像轮廓提取  

Extracting contours from image based on early cognitive vision system

作　　者：白蕾[1] 王聪慧[1] 蒋超 Bai Lei;Wang Conghui;Jiang Chao(Department of Electrical Engineering,Shanxi Polytechnic Institute,Xianyang 712000,China)

机构地区：[1]陕西工业职业技术学院电气工程学院

出　　处：《国外电子测量技术》2019年第12期16-20,共5页Foreign Electronic Measurement Technology

基　　金：陕西工业职业技术学院院内科研项目（ZK19-10）资助

摘　　要：早期认知视觉是连接早期视觉和认知视觉的中间平台。与像素携带的弱的语义信息相比,提出了一种基于视觉基元的图像轮廓的表示方法。首先由早期认知视觉系统提取出图像的基元,基元是图片小块,有丰富的语义信息包括位置、方向、相位和三色值;通过基元的共色性和共线性得到图像的轮廓。实验结果表明,此方法能将图像的轮廓完好地提取出来,有效地保留了图像的必要信息并具有强抗噪性。由此得到的轮廓所具有的几何信息和表面信息能很好的用于物体模型的学习及未知物体的抓取等。The early cognitive vision system(ECV)is an intermediate stage that mediates information between the early vision stage,and the cognitive vision stage.Different with the weak semantic content carried by pixels,we present a novel representation of image contour,based on local symbolic descriptors that we call visual primitives.primitives are extracted by ECV from images.A primitive represents a small image patch in terms of position,orientation,phase and three color values,according to the co-linearity and co-colority between primitives,the contours of the image can be obtained.The experiment result shows the contours of image completely,that reflect the necessary information of the image and the method is insensitive to noise.The contours representation covers both geometric and appearance information that can be used in several applications,in particular in the area of object model learning and grasping unknown objects.

关 键 词：早期认知视觉 基元 共色性 共线性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]