基于视觉的航天电连接器的智能识别与装配引导

发布时间：2018-07-13 19:48

【摘要】：为提高操作空间狭小的精密航天产品的手工装配效率,提出一种基于视觉的航天电连接器的智能识别与装配引导方法。利用肤色特征与零件的轮廓特征提取原始图像中零件所在的区域。提取电连接器的斑点特征和尺度不变特征变换特征放入支持向量机的分类器,训练得到零件训练模型。借助Hough变换与零件训练模型实现了电连接器的在线分类识别,整个识别过程无需人工干预。通过预先构建完善的三维装配工艺信息模型,利用识别结果实时触发并调取对应零件的装配工艺信息,用于指导装配。实例验证表明,零件识别平均准确率达90%以上,单幅图片识别时间在2s内,能够满足在线识别精度和效率的要求。
[Abstract]:In order to improve the manual assembly efficiency of precision spaceflight products with narrow operating space, an intelligent identification and assembly guidance method for aerospace electrical connectors based on vision is proposed. The skin color feature and the contour feature of the part are used to extract the region of the part in the original image. The speckle feature and the scale invariant feature transform feature of the electrical connector are extracted and put into the classifier of the support vector machine and the part training model is obtained. With the help of Hough transform and part training model, the online classification and recognition of electrical connectors are realized, and no manual intervention is needed in the whole recognition process. By constructing a perfect 3D assembly process information model in advance, the identification result is used to trigger the assembly process information of the corresponding parts in real time and is used to guide the assembly. The example shows that the average accuracy of part recognition is more than 90%, and the recognition time of single image is within 2 s, which can meet the requirements of on-line recognition accuracy and efficiency.
【作者单位】：上海交通大学机械与动力工程学院智能制造与信息工程研究所;上海航天精密机械研究所;
【基金】：上海航天先进技术联合研究中心资助项目(UACAST2015-17)~~
【分类号】：TP391.41;V465

【相似文献】