Multi-Network Collaboration for Three-Stage Armature Defect Detection
The armature is an important component of a micro motor,and its surface quality directly affects the operation of the motor.Currently,the accuracy of armature surface quality detection is low,and it is dif-ficult to control the features.In order to improve information capture efficiency and achieve precise classifi-cation and identification of armatures′defects,article proposes a multi-network collaborative three-stage ar-mature detection method.Firstly,MobileNetV2 is used to classify the integrity of copper wire and variable resistance areas in armature images.Then,the detected images are processed using YOLOv4 for feature rec-ognition.Finally,pattern recognition is performed based on the recognized detection results.Experimental results show that this method can accurately and precisely classify armature defects.
万方数据
维普
