| 王玉豪， 方贵盛.基于DAGSVM和决策树的电气草图符号识别[J].轻工机械,2017,35(4):56-59 |
| 基于DAGSVM和决策树的电气草图符号识别 |
| Electrical Sketch Symbol Recognition Based on DAGSVM and Decision Tree |
| |
| DOI：10.3969/j.issn.1005 2895.2017.04.012 |
| 中文关键词: 符号识别 决策树 有向无环图支持向量机（DAGSVM） 笔画特征 |
| 英文关键词:symbol recognition decision tree DAGSVM（directed acyclic graph support vector machine） stroke feature |
| 基金项目:浙江省自然科学基金项目资助（LY13F020032） |
|
| 摘要点击次数: 1405 |
| 全文下载次数: 1593 |
| 中文摘要: |
| 针对电气行业设计人员构思初步电路方案徒手绘制电气草图时，存在的电气符号模糊性和输入的随意性等缺点，提出基于有向无环图支持向量机(DAGSVM)和决策树的组合符号识别模型。提出了改进的有向无环图支持向量机多分类算法用于基本符号识别和组合符号识别；在组合符号预分类过程中引入决策树算法用来减少分类器的数量。应用结果表明：应用该模型的系统不仅能够有效识别出各种复杂的手绘电气符号，且有效地降低了识别的计算成本。该系统能帮助电气工程师将设计思想快速、清晰地转换为电气工程图，提高了设计效率。 |
| 英文摘要: |
| Due to some shortcomings of drawing freehand electrical sketches when electrical engineers conceive preliminary design schemes，such as poor effects of fuzziness of electrical symbols and randomness of inputs，etc，a combined symbol recognition model based on DAGSVM and decision tree was proposed．The decision tree was introduced to reduce the number of classifiers in the pre classification process. Application results show that the proposed model can not only effectively identify the various kinds of complex hand drawn electrical symbols, but also largely reduce the computational costs. The system can help the electrical engineers convert their design ideas to electrical drawings fast and clearly. Meanwhile, the design efficiency is improved. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
