基于改进型YOLO v3的蔬菜识别算法  

Vegetable Recognition Algorithm Based on Improved YOLOv3

作　　者：魏宏彬 张端金[1] 杜广明 肖文福 WEI Hongbin;ZHANG Duanjin;DU Guangming;XIAO Wenfu(School of Information Engineering,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]郑州大学信息工程学院,河南郑州450001

出　　处：《郑州大学学报（工学版）》2020年第2期7-12,31,共7页Journal of Zhengzhou University（Engineering Science）

基　　金：国家自然科学基金资助项目(61471323)。

摘　　要：针对超市的散装蔬菜区排队称重问题(称重设备能够自动识别蔬菜种类将有效地提高超市的运行效率),提出一种基于改进型YOLOv3的蔬菜识别方法。首先,利用高清摄像头以及网络爬虫技术采集蔬菜图片;其次,通过K-means聚类分析得到15组适应于蔬菜数据集的先验框;再次,采用一种新的边界框回归损失函数DIoU来提高检测任务的精度;最后,因蔬菜数据集中的大目标较多,通过增强特征提取网络,获取5组不同尺度的特征构成特征金字塔从而实现蔬菜识别任务。改进型YOLOv3算法在测试集上的平均精度mAP达到93.2%,识别速度是35 f·s^-1。该方法在保证实时检测目标的同时提升了识别的平均精度。The queuing and weighing problem was common in bulk vegetable area of supermarket.If weighing equipment could automatically recognize vegetable,it would effectively improve the operational efficiency of supermarket.Therefore,a vegetable recognition method based on improved YOLOv3 was proposed.Firstly,vegetable pictures were collected by using high-definition camera and web crawler technology.Secondly,15 groups of anchors suitable for vegetable datasets were obtained by K-means clustering analysis.Thirdly,a new bounding box regression loss function DIoU was proposed to improve the precision of detection task.Finally,as there were many large objects in vegetable datasets,5 groups of feature pyramids with different scales were obtained by enhancing feature extraction network to realize vegetable detection task.The mAP of the improved YOLOv3 algorithm on the test dataset was 93.2%,and the recognition rate was 35 fps.This method improved the recognition of mAP while guaranteeing real-time object detection.

关 键 词：蔬菜识别 K-MEANS 卷积神经网络 特征金字塔 YOLOv3 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]