Development and evaluation of a platform for the automatic extraction and processing of PCB information from schematics and placement files

The invention of the Printed Circuit Board (PCB) led to different breakthroughs in the electronic world. PCBs are the crucial basis for virtually every electronic device and find application in several fields, ranging from medical to automotive and IoT to aerospace. An extensive number of components find a place on the topmost or the bottom layer, and they are connected following an intricate tangle of connections. Two types of files are essential for understanding how a PCB works: placement (describes the layout of components on one layer) and schematic (reports the connections and the pins for each component). Manually testing a PCB requires looking for a specific component or connection in a placement or schematic file, which in turn requires the tester to look alternatively between the board and the manual. Avoiding looking at the manual for each component and connection would result in an acceleration of the testing phase, apart from reducing human errors: this can be achieved by synthesising all the information required for testing in an Augmented Reality (AR) application. The idea is to have a virtual layout that superimposes the board itself, which allows the tester to verify the correct functioning of the components without needing to inspect the manual. Certainly, some information processing is needed to reach the goal: by just giving as input the placement and the schematic files of interest, the proposed program automatically extracts the information required for the AR testing phase. The application relies on text extraction tools, graphical user interface packages and image processing libraries. The user has to upload the manual and select the placement or schematic files to analyse: then an ad hoc pipeline is launched distinguishing the operations based on the type of file currently processed. Some manual corrections may be required since components are obtained from placement files using an edge detection algorithm, that may encounter troubles and limitations. At the end of the process, the extracted and elaborated information is written in some textual files respecting a format suitable for the AR application, keeping track of only those components appearing in placement and schematic files of interest.