基于机器视觉的分切刀具豁口检测

Research on cutting tool gap detection based on machine vision

新能源电池极片毛刺是导致电池短路爆炸的重要因素之一,其主要是生产过程中使用带有豁口的刀具进行分切导致,因此本文研究分切刀具豁口的高精度、快速检测算法,用于检测分切刀具豁口。首先,本文依托于一套分切刀具图像采集装置采集分切刀具边缘显微图像;其次,在刀具边缘提取方面,提出了一种改进的Zernike矩亚像素边缘算子,实现亚像素级的边缘提取,有效保留分切刀具刃口边缘特征;最后,在刀具豁口检测方面,提出了一种基于四次Hermite插值的豁口检测算子,通过预推理分切刀具初始边缘,实现分切刀具豁口检测。将本文提出的算法与PreNet检测算法和基于三次样条插值的检测算法进行对比实验,实验结果表明:本文算法实现了亚像素级别分切刀具边缘提取,速度更快,分切刀具豁口检测准确度更高。

New energy battery electrode burr is one of the important factors leading to battery short-circuit explosion,which is mainly caused by the use of cutting tools with gaps in the production process.Therefore,this paper studies a high-precision and fast detection method for cutting tool gaps,which is used to detect cutting tool gaps.Firstly,this paper collects the micro image of the cutting tool edge based on a set of cutting tool image acquisition device.Then,in the aspect of tool edge extraction,an improved Zernike moment sub-pixel edge operator is proposed,which can realize sub-pixel edge extraction and effectively preserve the edge characteristics of the cutting tool.Finally,in the aspect of cutting tool gap detection,a cutting tool gap detection operator based on quartic Hermite interpolation is proposed to detect the cutting tool gap by pre-reasoning the initial edge of the cutting tool.The proposed algorithm is compared with PreNet detection algorithm and detection algorithm based on cubic spline interpolation.The results show that the proposed algorithm achieves faster sub-pixel cutting tool edge extraction speed and higher accuracy of cutting tool gap detection.