| The main function of fabric samples stored by garment production enterprises is to display and promote fabric products,and promote the communication between garment enterprises and between garment enterprises and customers.At present,the management of fabric sample storage system mainly relies on manual to complete the sorting and information management of fabric samples.However,the workload of fabric sample storage management is large,the information is complicated,and it is very easy to appear the chaotic situation of fabric sample storage,which leads to the low management efficiency and high cost of fabric sample storage management.At present,the improvement and upgrading methods for high density material storage are mainly warehousing and sorting integration system integrating shelf shuttle technology and high density three-dimensional shelf technology.However,for a wide variety of different hanging fabric sample storage problems,the existing storage equipment can not meet the needs of high density storage and automatic sorting,so this paper analyzes the characteristics,practical problems and needs of hanging fabric sample storage,and studies the automatic identification and positioning system of hanging fabric sample.Firstly,by studying the storage characteristics and sorting needs of hanging fabric samples,the overall storage sorting scheme based on intelligent storage cabinet and visual identification positioning of fabric samples is determined.In order to reduce manual labor intensity and replace manual fabric sample sorting,robot combined with online vision method is selected to identify,locate and grab fabric samples.In order to achieve the goal of robot automatic sorting,and considering the characteristics of fabric samples are high-density storage,variety and different,the fabric sample storage cabinet is developed and designed.Since the vision can not directly read the information from the fabric sample itself,QR code technology is introduced as the information storage carrier to store the information of the corresponding fabric sample.In order to facilitate visual identification of QR code,fabric sample hooks are designed to post QR code of corresponding fabric samples for visual identification.Secondly,through the analysis of the automatic sorting scheme of suspended fabric samples determined in this paper,it can be seen that the target detection algorithm based on machine vision is very important for QR code recognition and positioning.In order to achieve the rapidity and accuracy of multi-QR code small target detection under the big screen,this paper proposes a QR code rapid recognition and positioning method based on deep convolutional neural network.The YOLO v4 target detection algorithm was improved by lightweight processing method and combined with the MobileNet v2 feature extraction network,so as to achieve the lightweight network model while maintaining the accuracy and rapidness of QR code detection,realizing the purpose of the target detection algorithm to detect the accuracy and rapidness of QR code as much as possible in the big picture.Finally,based on the above development and design of fabric sample storage cabinet,the research on the detection algorithm of accurate and rapid identification and positioning of QR code corresponding to fabric samples,and the 3D target positioning technology of RGB-D camera,the fabric sample identification and positioning system is built.The accuracy and stability of the hanging fabric sample storage and rapid identification and positioning method studied in this paper are verified by the analysis of the QR code detection and identification results.Therefore,the research of this project achieves the goal of automatic sorting of fabric sample storage and improves the information management level of fabric samples. |
PDF Full Text Request