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2 s内,能够满足在线识别精度和效率的要求.%To improve the manual assembly efficiency of aerospace products,a visual-based intelligent recognition and assembly guide method of aerospace electrical connectors was proposed.The area of parts in the original image was extracted with skin color feature and contour feature of electrical connector.The feature of electrical connector and SIFT feature were extracted into support vector machine classifier training to get the part training model.With Hough Transform (HT) and part training model,the on-line classification and identification of electrical connectors were realized.The whole process was automated.Through the pre-construction of three-dimensional assembly process information model,the corresponding parts of assembly process information was triggered and retrieved with identification results in real time to guide the assembly.The test result of the proposed method proved that the average accuracy rate of part recognition was more than 90％ and the recognition time of each image were in two seconds,which could meet the requirements of identification accuracy and efficiency.
展开▼
