Research On Classification And Recognition For Aerospace Electrical Connector Oriented To Assembly Guidance

Aerospace electrical connectors plays an important role as the part of precision aerospace products.With familiar shape,hard to distinguish and assembly process is complex.In addition,precision aerospace products on the equipment consistency requirements,and product development to the stereotypes assembly process needs to be adjusted several times.All those factors lead to high requirements of the skill as assembly workers,and the assembly process needs repeatedly verify the correctness of assembly,thus affecting the assembly quality and efficiency of aerospace products.In order to solve the above problems,this dissertation,based on the deep analysis of aerospace electrical connectors,starts with four aspects: image feature extraction and classification,character-based template matching recognition,image classification and recognition based on neural network,and assembly-oriented intelligent information guidance Carried out the corresponding technical research,the results are as follows:Through in-depth analysis of the image of the aerospace electrical connector,part features are extracted from the aspects of geometric features,structural features and regional features,and the aerospace electrical connectors are classified and identified based on these features.For each aerospace electrical connector corresponds to a unique character number,based on the template matching electrical connector number identification.Through reasonable attitude adjustment to obtain the required character images,through the single character segmentation and matching with the template get the correct number of electrical connectors.In order to improve the accuracy and speed of classification and recognition of aerospace electrical connectors,this paper proposes a network structure for classification and recognition of aerospace electrical connectors based on the classic convolutional neural network structure,and further applies Faster-RCNN to aerospace electrical connectors Target detection and feature recognition.Finally,by comparing the advantages and disadvantages of the above three recognition methods,this paper selects the neural network classification and recognition method.Constructing the assemble guidance information model based on aerospace electrical connector.A prototype system was developed.Experiments on aerospace electrical connectors show that the system can quickly and accurately identify the electrical connector model and guide the assembly operation in real time.