Research On Visual Positioning And Identification Technology Of Express Package Information Code

The rapid development of China’s express delivery industry has brought fierce competition for express delivery time.As a key link in the operation process of express delivery,the efficiency and accuracy of the extraction of logistics packaging information will affect the sorting cost,operational efficiency,service level of express delivery enterprises,as well as the market competitiveness and development of enterprises in the next few years.In order to effectively extract the information of the flow of express parcels,systematic thinking and novel algorithms are used to process the logistics packaging information,and the deep learning in visual technology is used to locate and identify the logistics packaging information,so as to reduce the fatigue of repetitive work and increase the efficiency of express sorting in the logistics sorting work.In order to better realize the full automation in intelligent logistics,first of all,the impact of different packaging types of express parcels on sorting is analyzed,and then the adjustment experiment of relevant software is completed to complete the reading of the tilted information code,and its practical reliability reaches 97.5%.Obtain the information of the logistics express bill number,so that the same type of express delivery at the same destination can be stacked after sorting,reduce the space occupied by the express delivery,and make the goods of the same type of packaging express be better protected.Secondly,the concept of deep learning in visual technology in sorting is summarized,and the traditional deep learning method is used to collect and process the relevant datasets of express parcels,and through the experiment of classifying and positioning after learning,the average accuracy value reaches 71.06%,and it is found that it is impossible to deal with small targets well.To achieve the classification and positioning of the relevant information of the express parcel,and to provide packaging for the positioning of the express parcel in subsequent sorting.Compared with the traditional vision technology,deep learning can not well identify and locate small targets,on the basis of the traditional algorithm,according to the characteristics of the object,the introduction of new deep learning algorithms and optimization.Through the introduction of a new 4-scale detection overall structure,the detection of small target express parcels is realized.Feature enhancement is carried out by introducing a multi-scale feature fusion structure.By introducing giouf regression,the positioning accuracy is improved.Finally,the dataset is collected and processed,and the average accuracy value of the optimized algorithm is 94.12%.Experimental results show that compared with the traditional detection algorithm,it has great advantages,and the detection rate meets the requirements of real-time,which has certain practicality and effectiveness.Finally,the logistics express parcel information extraction system is designed,which includes the positioning identification and classification function of logistics outer packaging,the positioning,recognition and decoding function of bar code and two-dimensional code in express packaging,the output function of express bill number and the classification output function of logistics outer packaging,and can accurately extract value information content from the input express image.System experiments show that the system can effectively locate and extract logistics express parcel information,and the system can be widely used in the classification of goods and other fields.