摘要
为了满足针对手机屏幕玻璃的尺寸和崩边的高精度检测要求,设计了一种基于最小二乘法的手机屏幕玻璃尺寸检测及崩边评价方法。通过采用线阵相机实时地采集高精度的图像,同时对目标边缘像素进行基于灰度的采样并用最小二乘法进行拟合比较,再对定位后的目标边界进行分段的二次拟合,最后建立模板以及误差响应函数对玻璃边缘与拟合结果进行高通滤波后的误差统计,利用层次凝聚聚类可以获得准确的崩边位置和崩边大小。为了验证方法的可行性进行了实验,实验结果表明,崩边的检测率达100%,误检率为2%,尺寸检测误差小于0.05%,尺寸检测误差约为0.04 mm。该方法能较准确地检测手机屏幕玻璃尺寸和崩边,同时具有精度高的优点,并满足了实时检测要求。
In order to meet the high-precision detection requirements of the phone screen glass size and collapse edge,this paper designs a method for phone screen glass size detection and edge collapse evaluation based on the least square method. Real-time high-precision images are acquired using the linear array camera,the target edge pixels are gray-based sampled and fitted using the least squares method. The target boundary after positioning is quadratic fitted again for the final error statistics of the glass edge and the fitting results are established by using the template and the response function. The exact edge position and the collapse size can be obtained by using hierarchical agglomerative clustering. Experiments were carried out to verify the feasibility of the method. Experimental results show that the detection rate of collapse edge is 100%,the false detection rate is 2%,the size detection error is less than 0. 05%,and the size detection error is about 0. 04 mm. The method can accurately detect the glass size and collapse edge of the mobile phone screen,with high precision and satisfies the real-time detection requirement.
出处
《电子测量与仪器学报》
CSCD
北大核心
2018年第2期92-96,共5页
Journal of Electronic Measurement and Instrumentation
基金
浙江省新苗人才计划(2017R409025)资助项目
关键词
屏幕检测
层次凝聚聚类
误差函数
最小二乘法
screen detection
hierarchical agglomerative clustering
error function
least square method
