Design And Implementation Of Traditional Chinese Medicine Delivery System Based On Image Features

Nowadays,automation equipment can be seen everywhere in our life,people's demand for automation equipment is also increasing.But the automation level of distribution system in TCM pharmacies is still low,which can't meet people's needs.In this paper,a Chinese medicine distribution system based on image features is designed for the distribution center of Chinese medicine pharmacy.By extracting image features of Chinese medicine bags,the identification of Chinese medicine bags is completed,the number of bags is identified and the bags are audited and identified respectively,so as to achieve the effect of automatic dispensing of drugs.After understanding the needs of the system,the overall scheme of the system is designed according to the analysis of the distribution center of traditional Chinese medicine pharmacy.The system is divided into upper computer and lower computer.In the lower computer part,the structure and work flow of the lower computer are designed,and the selection of PLC and servo motor is completed,and the communication mode and protocol are selected.In the upper computer part,the software part is mainly designed.The main function of the software is to identify the number of drug bags and to complete the inspection and identification of drug bags.Aiming at solving the problem of identifying the number of Chinese medicine bags,the problem of identifying the number of Chinese medicine bags is transformed into the number of connected areas in the recognition picture.Firstly,the image is grayed,denoised and binarized.And then the connected region of the image is calibrated and recognized.In the process of image binarization,the Dajin method is improved to solve the problem of losing image details under uneven illumination and optimize the effect of image binarization.Aiming at solving the problem of drug bag auditing,the feature matching method is used to complete auditing.The SIFT and SURF are analyzed and compared.It is concluded the matching effect of SURF is better.Because we identifies traditional Chinese medicine bags,there are many similarities between different bags.If we use the general SURF,there will be many mismatches.So the SURF is improved.The method is improved by secondary matching.First,the feature points are extracted by Euclidean distance,and then the feature points are filtered by cosine similarity.The improved algorithm has better matching effect.After that,template matching is performed to complete the recognition.Finally,the feasibility of the system is verified by test.