Research on K nearest neighbor identification of hand-drawn circuit diagram

Abstract

Aiming at the problem of accurately identifying each hand-painted electrical component in the hand-drawn circuit diagram, this paper proposes a region segmentation based on pixel number distribution, and a hand-painted electrical component recognition method based on KNN algorithm, the aim is to form a hand-drawn sketch recognition algorithm with better recognition indicators to improve the accuracy of recognition. In this paper, Python's graphic pixilation module is used as a pixel segmentation tool. Based on 35 types of standard electrical components, the database is established. The Euclidean distance and KNN algorithm are used to obtain the corresponding classification and output recognition results. By comparing with other methods, the better identification indicators achieved by the experiment verify the effectiveness of the method.