基于云技术的采摘机器人数据系统设计

打开文本图片集

中图分类号：S225.93；TP242.6 文献标识码：A 文章编号：2095-5553（2026）04-0163-08

Abstract：Harvesting robotsconstituteacritical componentof inteligentagricultural machineryand smart farming systems，with significantadvancements in fruit detection，recognition，and positioning grasping inrecent years.Despite theseadvancements，the aplication of harvesting robots inreal orchards remainsunderdeveloped.Challenges include adaptingtothedynamicandeverchanging unstructured environmentsoforchards，whichresults inlow harvesting success ratesandincreased timeconsumption.Additionally，current research predominantly focuses onisolatedisues such as enhancing target detectionaccuracy or optimizing picking path planning，rather than developing acomprehensive autonomous operation system atthe aplication level.A data system for harvesting robots has bee designed，leveraging cloud synergy to deploycomplex data procesing tasksliketarget detectionand localizationon the cloud.Thesystem is placed in a simulated environment for picking experiments，and the picking recognition rate is ≥92% ．The picking positioning error is between 3.3% and 5.4% .thereby improving harvesting precision and real-time performance. Environmentalparametersof theorchardanddatafrom the harvestingrobot'sdeviceend，includingreal-timeimages from depthcamerasandstatus parametersof roboticarms，are uploaded inreal-timetoanInternetof Things（IoT）platform via the MQTT protocol.Thissetup providesuserswith remotemonitoring capabilities through a web-based visual interface.During practical testing，this interface demonstrateddatacompletenessandreal-timeperformanceof various sensors，enabling accurate alerts for unexpected anomalies.

Keywords：harvesting robot； cloud collaboration；fruit detection；MQTT protocol;data processing

0 引言

智慧农业将传感器技术、物联网技术、通信技术系统应用到农业系统解决方案中，利用相关传感器和信息管理系统可以实现精准监测和农机智能化管理[1，2]。（剩余11526字）