Machine Vision for Device Tracking in a Smart Manufacturing Environment Based on Augmented Reality

Abstract

In a controlled network environment, such as the smart indoor manufacturing environment, the device identification and detection of components is challenging without prior knowledge of the design and implementation process. Thus, the concept of device identification for diagnosis and equipment maintenance by means of markerless augmented reality (AR) merits investigation. AR, when coupled with machine vision, caters for obtaining real-time device information regarding the position and features of the robotic elements within indoor manufacturing plants. Thus, this article proposes an efficient machine vision model to detect and identify devices based within a manufacturing plant, with the aid of AR for extending the device operational details. This offers an alternative solution in the absence of user built-in maps for the calculation of device positions based on uncertainties of the exact locations. To achieve this, a two-part validation is conducted involving (1) device recognition based on position and (2) Data integration to A Supervisory Control and Data Acquisition (SCADA) model developed in National Instruments Labview. The findings demonstrate that the AR application can detect devices within the manufacturing plant without the need for alteration. The results also indicate that the application can be integrated into a SCADA model without the need to alter the application, provided that the array index is the same. Only when the array index differs are alterations necessary for utilising the AR application.