摘 要:砝码智能化检定有助于提高砝码检定效率、增强检测数据可溯源性。文章从 20 kg 砝码组智能化检定系统整体设计入手,提出依靠三维堆放 20 kg 砝码组深度学习识别分割技术、多模态传感定位技术配合设计的软硬件平台,实现日光灯照明下三维堆放 20 kg 砝码组的智能检定。通过模拟搭建三维堆放砝码组开展系统测试试验,验证了智能检定系统的有效性、准确性,单个砝码实例检定时间约为 42 s,可以满足实际 20 kg 砝码组检定过程需要,实现“机器代人”,大幅提高 20 kg 砝码组检定效率。
关键词:20 kg 砝码组;深度学习;多模态传感;智能化平台
DOI:10.19850/j.cnki.2096-4706.2021.12.001
基金项目:国家市场监督管理总局科技计 划项目(2019MK086)
中图分类号:TP391.4;TH715 文献标识码:A 文章编号:2096-4706(2021)12-0001-05
Design of an Intelligent Verification System for 20 kg Weight Set
MA Jian1 , ZHAO Di 2 , GUO Linlin1 , LIU Guixiong2
(1.Guangzhou Institute of Measurement and Testing Technology, Guangzhou 510030, China; 2.School of Mechanical & Automotive Engineering, South China University of Technology, Guangzhou 510641, China)
Abstract: The intelligent verification of weights is helpful to improve the verification efficiency of weights and enhance the traceability of test data. Starting with the overall design of the intelligent verification system of 20 kg weight set, this paper proposes a software and hardware platform based on the deep learning, identification and segmentation technology of three-dimensional stacking 20 kg weight set and the multi-modal sensing and positioning technology, so as to realize the intelligent verification of three-dimensional stacking 20 kg weight set under the illumination of fluorescent lamp. The validity and accuracy of the intelligent verification system are verified by simulating the construction of three-dimensional stacking weight set and carrying out system test. The verification time of a single weight example is about 42 s, which can meet the needs of the actual verification process of 20 kg weight set, realize "machine replace people" and greatly improve the verification efficiency of 20 kg weight group.
Keywords: 20 kg weight set; deep learning; multi-modal sensing; intelligent platform
参考文献:
[1] 全国量和单位标准化技术委员会 . 国际单位制及其应用: GB 3100-1993 [S]. 北京:中国标准出版社,1993.
[2] 王益飞,杨圆圆 . 公斤组砝码自动检定系统 [J]. 计量技术, 2017(7):61-64.
[3] UIJLINGS J R R,SANDE K E A,GEVERS T,et al. Selective Search for Object Recognition [J]. International Journal of Computer Vision,2013,104(2):154-171.
[4] JIANG J J,GUI X Q. Target Detection and Recognition Method of Farming Machine Based on Machine Vision [C]//Proceedings of the 2015 4th National Conference on Electrical,Electronics and Computer Engineering.Paris:Atlantis Press,2015:1385-1389.
[5] 高云,郭继亮,黎煊,等 . 基于深度学习的群猪图像实例 分割方法 [J]. 农业机械学报,2019,50(4):179-187.
[6] 彭秋辰,宋亦旭 . 基于 Mask R-CNN 的物体识别和定位 [J]. 清华大学学报(自然科学版),2019,59(2):135-141.
[7] OCER N E,KAPLAN G,ERDEM F,et al. Tree extraction from multi-scale UAV images using Mask R-CNN with FPN [J].Remote Sensing Letters,2020,11(9):847-856.
[8] 段勇,裴明涛 . 基于多 RGBD 摄像机的动态场景实时三维 重建系统 [J]. 北京理工大学学报,2014,34(11):1157-1162.
[9] 曲磊,王康如,陈利利,等 . 基于 RGBD 图像和卷积神经 网络的快速道路检测 [J]. 光学学报,2017,37(10):124-132.
[10] ZOU C H,GUO R Q,LI Z Z,et al. Complete 3D Scene Parsing from an RGBD Image [J].International Journal of Computer Vision,2019,127(2):143-162.
[11] REIS D,WELFER D,CUADROS D S L,et al. Mobile Robot Navigation Using an Object Recognition Software with RGBD Images and the YOLO Algorithm [J].Applied Artificial Intelligence, 2019,33(14):1290-1305.
作者简介:马健(1980—),男,汉族,广东广州人,高级工 程师,硕士,研究方向:衡器计量技术研究。