| 基于共生矩阵和颜色特征的泡罩包装缺陷检测 |
| |
| 引用本文： | 龙永红,吴敏.基于共生矩阵和颜色特征的泡罩包装缺陷检测[J].湖南农业大学学报(自然科学版),2005,32(6). |
| |
| 作者姓名： | 龙永红 吴敏 |
| |
| 作者单位： | [1]中南大学信息科学与工程学院,湖南长沙410083 [2]株洲工学院电气工程系,湖南株洲412008 |
| |
| 基金项目： | 国家自然科学基金资助项目（60425310） |
| |
| 摘 要： | 提出了一种基于视觉的泡罩包装缺陷检测方法．首先将表面图像分成泡罩区和底板区两部：令．在泡罩区先采用基于HIS颜色空间的合成颜色特征矢量提取出药片边缘，计算药片属性，然后采用改进的闵可夫斯基距离法进行药片颜色匹配．底板区通过提取基于共生矩阵的纹理特征，采用反向传播神经网络分类器进行缺陷识别，实验证明：泡罩区采用的算法不仅药片边缘定位精度高，而且可有效检测药片的污物、色斑等缺陷；底板区采用的识别算法能够检测各种纹理的复杂缺陷．
|
| 关 键 词： | 泡罩包装 缺陷检测 颜色特征 共生矩阵 神经网络分类器 |
Automatic Visual Inspection for Foamed Mask Packing Products Based on Co-occurrence Matrix and Color Features |
| |
| LONG Yong-hong,WU Min.Automatic Visual Inspection for Foamed Mask Packing Products Based on Co-occurrence Matrix and Color Features[J].Journal of Hunan Agricultural University,2005,32(6). |
| |
| Authors: | LONG Yong-hong WU Min |
| |
| Institution: | 1. College of Information Science and Engineering, Central South Univ, Changsha,Hunan 410083,China; 2. Dept of Electrical Engineering, Zhuzhou Institute of Technology, Zhuzhou,Hunan 412008, China |
| |
| Abstract: | This paper presented a new system based on vision for defect inspection of foamed mask packing products.Surface image was divided into two parts: foamed mask region and soleplate region.In the foamed mask region,a synthesized color feature vector based on |
| |
| Keywords: | foamed mask packing defects inspection color features co-occurrence matrix neural network classifier |
|
| 点击此处可从《湖南农业大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《湖南农业大学学报(自然科学版)》下载免费的PDF全文 |
