High-precision dimensional measurement for digital crown part of smartwatch from microscopic image

Digital crown is a precision part used for human-computer interaction in the smartwatch. The diameter of the digital crown determines its grade. Currently, mainly used Coordinate Measuring Machine(CMM) to measure the diameter of digital crown, but the measurement method has the problem of high cost and low efficiency. To address the problem, we propose a high-precision dimensional measurement system for digital crown part of smartwatch. The system consists of four modules: microscopic image acquisition module, central host, machine vision software and PLC control module. The image of digital crown is captured by the microscopic image acquisition module and sent to the software through the central host. In the machine vision software, we propose a defect detection algorithm using ResNet to filter defective digital crown in real time, and we design a circle detection algorithm based on edge point tracking to measure the diameter of digital crown. The measurement result is sent to the PLC control module through wireless transmission and control the mechanical suction cups to place the part. Finally, we create a dataset to test the performance of the system, the defect detection model achieves an accuracy of 97.33% and a runtime of 47ms, the dimensional measurement algorithm achieves a maximum error of 0.003878mm, a grading accuracy of 95.29%, and a measurement time of 465ms.