基于机器视觉的LTE模块编带系统

LTE Module Taping System Based on Machine Vision

针对LTE(终端高速无线通讯标准)模块屏蔽盖表面镭雕字符信息提取速度和准确度较低,视觉检测与编带包装无法同时进行的问题,设计了基于机器视觉的LTE模块编带系统。通过中值滤波与灰度变换改善图像质量,形状匹配完成目标定位,使用MLP(多层感知器)分类网络与特征向量融合特相似度完成字符识别。经现场生产测试,系统字符识别速度达到0.0417 s/片,识别准确率达到97.3%,识别效率和准确率明显高于传统人工检测,系统在满足实际生产要求的同时,提高了LTE模块生产的效率和自动化程度。

Aiming at the problem that the extraction speed and accuracy of the laser-engraved character information on the surface of the LTE(terminal high-speed wireless communication standard)module shielding cover is low,and the visual inspection and taping packaging cannot be performed at the same time,a machine vision-based LTE module taping system was designed.The image quality was improved by median filtering and grayscale transformation,shape matching was used to complete target positioning,and MLP(multilayer perceptron)classification network and feature vector fusion similarity were used to complete character recognition.After on-site production test,the system character recognition speed reaches 0.0417 s/piece,and the recognition accuracy rate reaches 97.3%.The recognition efficiency and accuracy rate are significantly higher than traditional manual inspection.While meeting actual production requirements,the system improves the efficiency of LTE module production.and the degree of automation.