教授 博士生导师 硕士生导师
主要任职: 机械工程学院院长、党委副书记
性别: 男
毕业院校: 大连理工大学
学位: 博士
所在单位: 机械工程学院
学科: 机械电子工程. 测试计量技术及仪器. 精密仪器及机械
办公地点: 辽宁省大连市大连理工大学机械工程学院知方楼5027
联系方式: 辽宁省大连市大连理工大学机械工程学院,116023
电子邮箱: lw2007@dlut.edu.cn
开通时间: ..
最后更新时间: ..
点击次数:
论文类型: 期刊论文
发表时间: 2011-02-01
发表刊物: JOURNAL OF MATERIALS PROCESSING TECHNOLOGY
收录刊物: SCIE、EI
卷号: 211
期号: 2
页面范围: 237-244
ISSN号: 0924-0136
关键字: Dimensional measurement; Stereo vision; Feature points; Center points; Light stripe
摘要: Non-contact dimensional measurement of hot forgings is desirable for process control and product quality. The stereo vision method has the potential to detect the dimensions of hot forgings. However, it is necessary for the stereo vision method to obtain clear image, and to extract and match rapidly feature points of hot forgings. The spectrum selective method proposed has improved the image quality of hot parts. In this paper, based on the spectrum selective method, a fast method with the projected light stripes is presented for dimensional measurement of hot forgings. Firstly, the light stripes on the hot forging are resolved from the image by the proposed modified Gaussian model. Then, a fast sub-pixel extraction algorithm of center points of the light stripes is designed, especially for the situation of small gray variation in the image of hot forgings. Finally, the endpoints of the light stripes are employed to rapidly calculate the dimensions of hot forgings. Experiments on measuring the dimensions of hot forgings in the workshop are conducted to testify the proposed method. The experimental results indicate that the measurement method is efficient. Besides, an additional experiment in the laboratory verifies the accuracy of the presented method. Crown Copyright (C) 2010 Published by Elsevier B.V. All rights reserved.