基于YOLOv3的中医处方饮片自动检测研究与实现  

Research and implementation of automatic detection of the decoction pieces in traditional Chinese medicine prescription based on YOLOv3

作　　者：邱佳瑜 王艳[1] 吴浩忠[1] 韩爱庆[1] 翟兴[1] 唐燕[1] QIU Jia-yu;WANG Yan;WU Hao-zhong;HAN Ai-qing;ZHAI Xing;TANG Yan(Beijing University of Traditional Chinese Medicine,Beijing 100029,China)

机构地区：[1]北京中医药大学,北京100029

出　　处：《信息技术》2024年第3期35-42,48,共9页Information Technology

基　　金：2020年第二批教育部协同育人项目(BUCM-2020-CXY-19)。

摘　　要：随着中医药的普及,中药饮片及中医处方的智能辨识尤为重要。文中训练中医处方饮片自动检测模型,辅助药房自动核检、智能辨识中药饮片。以小柴胡汤加减方为例,构建中医处方饮片检测数据集,利用YOLOv3模型对数据集进行训练。YOLOv3模型收敛速度较快,测试集mAP为99.16%,平均FPS为10.11。YOLOv3模型对中医处方数据集有较好的检测效果,可以推广应用在处方药材的智能识别检测和药材鉴别等领域。该研究为中药材的识别及中医处方饮片的智能检测提供了一种新方法。With the popularization of traditional Chinese medicine,the intelligent identification of traditional Chinese decoction pieces and prescriptions is particularly important.In this paper,we train the automatic detection model of traditional Chinese medicine prescriptions decoction pieces to assist the automatic verification and intelligent identification of traditional Chinese medicine decoction pieces.Taking the addition and subtraction of Xiaochaihu Decoction as an example,the test data set of decoction pieces of traditional Chinese medicine prescriptions is constructed,and the YOLOv3 model is used to train the data set.The convergence speed of YOLOv3 model is fast.The test set mAP is 99.16%and the average FPS is 10.11.As a result,the Yolov3 model has a good detection effect on the data set of traditional Chinese medicine prescriptions,and can be widely used in the fields of intelligent identification and detection of prescription medicinal materials and identification of medicinal materials.This paper provides a new method for the identification of traditional Chinese medicine and the intelligent detection of traditional Chinese medicine prescriptions decoction pieces.

关 键 词：深度学习 目标检测 中医药处方 YOLOv3 中药材 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R288[自动化与计算机技术—计算机科学与技术]