Key Technology Research On Color Printing Of Takeout Packaging Boxes Based On Image Recognition

With the rapid development of the takeaway industry,consumers are increasingly concerned about the printing quality of takeaway packaging boxes.Traditional manual inspection methods are inefficient,subjective and have high leakage rates,and lack an automated inspection technology.The emergence of image recognition technology has brought a new solution to traditional printing defect detection,and has received key attention from printing companies due to the simplicity,efficiency and accuracy of image recognition technology.Based on the above background,this study is based on image recognition technology to investigate the colour defects that occur in the printing process of takeaway boxes.Traditional target detection algorithms are analysed according to the key problems of small target colour defects and three-dimensional printing colour defects that occur in the printing process of takeaway boxes.The YOLOv5 algorithm is used as a key technology to improve the current key problems in the printing of takeaway boxes,combined with the actual production of enterprises.To address the problem that the YOLOv5 algorithm is poor at detecting small target defects in the process of colour detection for takeaway box printing,and at the same time cannot accurately identify three-dimensional printing defects that may occur due to the embossing process.This study firstly uses histogram equalisation,morphological processing,image segmentation techniques and colour space conversion techniques to process the datasets in the image pre-processing stage,and secondly proposes to improve the traditional YOLOv5 algorithm using K-Means algorithm and Point Net++ algorithm.The anchor frames appearing in the printing process are clustered to enhance the recognition of small targets,while the Point Net++ algorithm is used to reconstruct the 2D image in 3D,enhancing the defect features to enable the recognition of three-dimensional printing defects,enabling better detection of small targets and three-dimensional printing defects.The robustness of the algorithm is improved.Finally,experimental validation of the data set shows that the proposed algorithm for small target colour defects and three-dimensional print defects in the printing process of takeaway boxes is more robust than traditional high-accuracy target detection in terms of overall performance evaluation metrics,and is more applicable and economical.The results of this study have important practical application value for automated improvement of print quality inspection of takeaway boxes,and can provide efficient and accurate print quality inspection solutions for the takeaway industry.The realisation of real-time monitoring and remote management of printing quality inspection provides a solution to realise the intelligence and informationisation of printing quality management.