Design of engineering drawing recognition system based on Yolo V4
Yanpei Luo, Changyuan Yu, Jiang Zheng, Yao Ding
Abstract
In order to realize the detection of devices in engineering drawings, this paper proposes an engineering drawing recognition method based on YOLOv4 algorithm, which is applied to the detection of component targets under circuit diagrams. Traditional drawing auditing is mostly done by manual means. With the development of information technology, applying image technology and pattern recognition to drawing auditing is a hot spot for the application of artificial intelligence in drawing auditing. YOLO is a simple and uncomplicated algorithm that can make predictions based on the whole picture information and learn the picture features more generally. The components (capacitors, resistors, inductors and other basic components) in the circuit diagram are identified, and their number and location are indicated.