A circuit board component detection method based on yolo9000 algorithm

Abstract

The invention discloses a circuit board component detection method based on a YOLO9000 algorithm. The circuit board component detection method comprises the following steps: S1, constructing a database comprising circuit board pictures; S2, training the YOLO9000 algorithm model by using a database to obtain a trained YOLO9000 algorithm model; S3, acquiring a to-be-detected circuit board picture; S4, detecting the circuit board picture to be detected by using the trained YOLO9000 algorithm model; S4, outputting a result. According to the invention, a YOLO9000 algorithm model is trained; the model type identification of the electronic component can be carried out; the electronic component is positioned on the image position, so that the model category, position and welding error of the electronic component are detected, and meanwhile, the direction identification problem is converted into the classification problem due to the welding direction of the polar electronic component on the circuit board, so that the direction identification of the polar electronic component is realized.

Description

technical field [0001] The invention relates to the field of computer vision, and more specifically, to a circuit board component detection method based on the YOLO9000 algorithm. Background technique [0002] In the industrial production of electronic products, testing after soldering electronic components is an essential link, but the current testing method is to use electronic instruments for testing, which wastes manpower and material resources to a certain extent. Moreover, soldering reverse polarity electronic components will cause greater losses during power-on detection. Deep Learning Since Krizhevsky et al. proposed a deep convolutional neural network (DCNN) called AlexNet in 2012, it achieved record-breaking image classification accuracy in the large-scale visual recognition challenge (ILSRVC). Subsequently, deep learning has developed rapidly in the field of target detection, and the YOLO series of algorithms continue to emerge. YOLO9000 has excellent performanc...

AI Technical Summary

Problems solved by technology

[0003] 1. Cannot identify the direction of polarized components;

[0004] 2. Features need to be extracted manually, and only simple classification can be realized, but complex classification cannot be realized;

[0005] 3. The components on the picture cannot be classified and positioned at the same time

Images