引入同现度概念的零件辅助视觉加工优化方法研究

Auxiliary Visual Parts Processing Optimization Method Research Based on the Concept of Degree of Co-occurrence

车床在对零件进行切削加工时,零件全程都处于高速旋转运动中。由于存在零件塑性变形现象和刀具积屑瘤现象,使得零件在加工过程中容易出现缺陷部位,造成零件质量下降,因此需要及时检出零件缺陷部位并调整机床,避免缺陷部位进一步扩大。传统的检测方法误差大,不能对零件缺陷部位进行准确定位,降低了对零件加工过程检测的准确性。为此,提出一种基于辅助视觉的零件加工过程检测方法。对加工过程中零件缺陷部位的辅助视觉图像出现同现的频率进行统计,计算缺陷部位图像在所有图像中的同现度,通过计算得到对应图像的权重,实现缺陷部位关键帧图像定位,利用缺陷图像中心点的位置建立数学模型,实现对零件缺陷部位的检测。实验结果表明,利用辅助视觉算法在零件加工过程中能够快速准确的对零件的缺陷部位进行检测。

Lathe in machining of parts, all the parts are in high speed rotation. Plastic deformation phenomenon due to parts and tool devolop tumor phenomenon, make parts prone to defects in the process of machining parts, cause the decrease of the quality of the parts, so need to check out the parts defective parts in time and adjust machine, avoid the defect part is widening. Traditional test method of error is big, can＇t to accurate positioning of parts defective parts, reduce the detection accuracy of the parts processing process. For this, put forward a kind of based on auxiliary parts processing process of visual detection methods. Defective parts in the process of machining parts of auxiliary visual image co-occurrence frequency statistics, calculating defect parts image in all images, co-occurrence degree, is obtained by computing the weight of corresponding image, realize the defect part key frames image positioning, mathematical model was established based on the location of the defect image center, realize the detection of defective parts of the parts. The experimental results show that the use of auxiliary vision algorithms in the process of parts processing can fast accurate to test the components of the defective parts.