摘要
针对自动引导车上的图像数字在识别时受现场环境影响大的缺陷,提出了一种基于特征融合的自动引导车图像数字识别方法,它将各工位图像数字归一化处理后,提取灰度信息、改进的穿越线特征等特征量进行特征融合计算,并输出最后结果。现场运行实验表明:在有噪声的情况下,该方法不仅能够提高图像数字的识别率,并且对由环境影响所导致的图像数字局部污染与残缺具有很好的鲁棒性。
Aiming at the deficiency of image number characters captured by AGV which is affected by scene environment, a method of image number recognition based characters fusion was brought forward. Firstly, image numbers of the groups were divided as one. Secondly, the characters such as ash degree information and across line improved were extracted and managed of fusion. Finally, the result of image number was worked out, The experiment course indicate that, the method not only can greatly improve the accuracy in recognition of image number, but also has a perfect identification effect and extent. It has a good robustness.
出处
《机电工程》
CAS
2007年第4期4-6,共3页
Journal of Mechanical & Electrical Engineering
基金
国家自然科学基金资助项目(50505045)
关键词
自动引导车
图像数字
特征提取
特征融合
automatic guided vehicle （AGV）
image number
characters extraction
characters fusion
