An Automatic Digital Instrument Quality Inspection System Based on Machine Vision

With subjective influence, misdetection and missing detection are inevitable problems for traditional manual quality inspection of digital instruments. This study designs an automatic quality inspection system based on machine vision, which can be widely applied in five-component digital instruments with an arbitrary layout on a 6 (2*3)-instrument test platform. The system uses Hikvision industrial camera to collect images and Python to program. The system consists of the instrument display standard formulation part and inspection part. The inspection part is divided into five steps: image acquisition and segmentation, instrument model identification, keyframe extraction, image preprocessing, digital display panel recognition and detection. The paper uses the 7-point identification method to identify the state of the nixie tubes and uses the mean square deviation method to identify the states of the indicator lights. It has been proved that the system inspection can control the missing detection rate at 1.59%o ~3.70%o, which shows that the system can inspect the quality of digital instruments with an arbitrary layout automatically and accurately.