Circuit Component Detection in Offline Handdrawn Electrical/Electronic Circuit Diagram

Abstract

Detection of circuit components from handdrawn circuit image refers to the process of automatic separation of all the circuit components belonging to a circuit. In this work, we have devised an algorithm that accepts a hand-drawn circuit image as input and returns the circuit components marked with bounding box. We have proposed a method where we have used the morphological `close' operator to create a structural dissimilarity between the circuit components and the wires. Then we have used run length smoothing algorithm (RLSA) based technique to remove all the wires. Finally, we have used the spatial coordinates of the circuit components in the wire-removed circuit diagram to localize them in the actual circuit diagram. The proposed methodology is evaluated over a database consisting of 60 circuit images collected from the copies of engineering students. Considering the complexity of the research topic, the result thus obtained is encouraging.