上海海事大学港口物流数智化研究团队始终致力于推动港口智能化发展,融合机器视觉识别、数字孪生和无人机管控技术,构建基于孪生管控的港口物流安全监管体系,融合港口智能识别和时空孪生监控无人机智慧巡检系统,实现了事前预警、事中分析、事后回溯的全生命周期智能安监服务,对港口生产进行实时监管和控制,旨在提高安全生产管理的效率和质量,预防和减少安全事故的发生。
其中,港口机器视觉系列产品提供港口安全识别全时域智能解决方案,时空一体化无人机港口智能巡检系统通过全生命周期安监赋能,实现“一张图”式管控,最终实现港口作业从人工智能识别感知到认知决策全覆盖跃迁。
图1 平台架构
01
港口安全识别
港口安全识别涵盖港口设施、船舶作业、货物安全和人员安全等领域,通过机器视觉技术监控港口运营活动,预防港口作业安全事件,提高港口应急响应能力,确保港口安全高效运营。
图2 海大港口安全识别产品
团队研发的各识别算法模块年识别集装箱量过亿,所有模块平均识别率超过98%,部分模块识别率接近100%(人眼勉强可辨别)。
(1)人员安全帽反光衣识别
功能
检测港口作业区域人员是否穿戴安全帽和反光衣,当出现人员未按照要求穿戴时,触发警报并通知现场安全管理人员进行干预,减少因未遵循安全规定而导致的事故。
适用
码头前沿、堆场、施工位置等多个应用场景。
图3 人员安全帽反光衣识别
(2)船上作业安全绳检测
功能
对船上作业过程中未按规定系安全绳等行为检测,提前预警,以确保发生吊装意外时能够提供有效保护,提高船舶安全作业水平。
图4 船上作业安全绳检测
(3)轨道异物识别
功能
识别桥吊、轨道吊等设备轨道上是否有异物、是否有人员等情况,提前预警,为港口机械设备安全运营提供有力保障。
图5 轨道异物识别
(4)车辆速度检测
功能
检测港口车辆行驶速度,对超速车辆进行预警,预防和减少车辆超速引发的作业事故。
图6 车辆速度检测
(5)堆场堆垛检查
功能
实时检查堆场堆垛箱是否存在倾斜或损坏等问题,为堆场安全管理、风险评估和堆存优化提供支持。
图7 堆场堆垛检查
(6)堆场绑扎检查
功能
自动识别堆场不正确的绑扎或固定,支持在台风到达前检查堆场是否绑扎,避免货物在堆存过程中移位、倒塌或损坏,从而造成安全事故和财产损失。
图8 堆场绑扎检查
(7)危险区域人员识别
功能
识别港口高风险区域出现的人员,并给出预警信息,避免人员在这些区域出现生产事故。
图9 危险区域人员识别
港口在线机器视觉API平台体验版
访问链接:http://47.103.112.75:8501/user/login
用户名:cv_test1
密码:123456
02
时空一体化无人机港口智能巡检系统
团队进一步融合无人机管控、机器视觉识别和数字孪生技术,研发了港口智能识别和时空孪生监控无人机智慧巡检系统,实现了事前预警、事中分析、事后回溯的全生命周期巡检。
构建港口安全管理和运营预警监控体系,实现实时动态显示生产设备状态和告警信息、实时动态跟踪危险货物全作业流程、实时监控码头安全生产运营,智能判断潜在风险。
图10 区域温度异常预警(样例)
自主研发无人机控端APP智能设置无人机最优巡检路径,实现无人机动态巡航管理,视频数据实时回传和历史巡检任务回溯,解决了人员高空作业风险大、人工巡检成本高等问题。
图11 无人机巡检路径设置(样例)
全面采集和存储港口物联网实时数据和预警信息,融合孪生平台准确重现历史运营环境和风险态势,识别运营中的瓶颈和不足之处,从而发现潜在问题并优化未来安全管控策略。
SMU Port intelligent safety supervision product released: Port logistics security supervision based on twin control
Shanghai Maritime University port logistics digitization and intelligent research team has always been committed to promoting the intelligent development of the port. The team integrates machine vision recognition, digital twin and drone management and control technology to build an intelligent safety supervision and control system, integrates port intelligent identification and space-time twin monitoring UAV intelligent inspection system, realizes the full life cycle intelligent safety monitoring service of advance warning, in-process analysis, and Ex-post facto, and carries out real-time supervision and control of port production. It aims to improve the efficiency and quality of safety production management, prevent and reduce the occurrence of safety accidents.
Among them, the port machine vision series products provide Port Security Identification, Time-space Integration Unmanned Aerial Vehicle Port Intelligent Inspection System realize "one map" control through the whole life cycle safety supervision, and finally realize the port operation from artificial intelligence recognition perception to cognitive decision-making full coverage.
Figure 1 Platform architecture
01
Port Security Identification
Port security identification covers the areas of port facilities, ship operations, cargo safety and personnel safety. It supports the monitoring of port operation activities and the prevention of port operation security incidents, improves port emergency response capabilities, and ensures safe and efficient port operations.
Figure 2 Port Security Identification
The recognition algorithm modules developed by the team identify more than 100 million containers a year, the average recognition rate of all modules exceeds 98%, and the recognition rate of some modules is close to 100% (barely identifiable by the human eye).
(1)Personnel Safety Helmet and Safety Vest Identification
Function
Detect whether personnel in the port operation area are wearing safety helmets and safety vests. When personnel fail to wear them as required, trigger an alert and notify on-site safety management personnel to intervene, thereby reducing accidents caused by failure to follow safety regulations.
Application
quayside, yard, construction location and other application scenarios.
Figure 3 Personnel Safety Helmet and Safety Vest Identification result
(2)Ship Operation Safety Rope Detection
Function
The identification detects behaviors such as not wearing safety ropes during shipboard operation and gives early warning to improve the level of ship safety operation.
Figure 4 Ship Operation Safety Rope Detection result
(3)Track Foreign Body Identification
Function
The identification identifies whether there are foreign Bodies or personnel on the track of equipment such as quay crane and yard crane, and gives early warning in advance to provide guarantee for the safe operation of port equipment.
Figure 5 Track Foreign Body Identification result
(4)Vehicle Speed Detection
Function
The identification detects the speed of port vehicles, gives early warning to speeding vehicles, and prevents and reduces operational accidents caused by speeding vehicles.
Figure 6 Vehicle Speed Detection result
(5)Yard Stacking Inspection
Function
The identification checks in real time whether there are problems such as tilting or damage in the yard container, providing support for the safety management of the yard, risk assessment and storage optimization.
Figure 7 Yard Stacking Inspection result
(6)Yard Banding Inspection
Function
The identification automatically identifies the incorrect banding or fixing of the storage yard, and supports checking whether the storage yard is banding before the arrival of the typhoon to avoid the displacement, collapse or damage of the goods during the storage process, resulting in safety accidents and property losses.
Figure 8 Yard Banding Inspection result
(7)Personnel Identification in Dangerous Areas
Function
The identification identifies the personnel present in the high-risk areas of the port, and gives early warning information to avoid production accidents in these areas.
Figure 9 Personnel Identification in Dangerous Areas
Port online machine vision API platform experience version
Website: http://47.103.112.75:8501?locale=en-US
User name: cv_test1
Password: 123456
02
Time-space Integration Unmanned Aerial Vehicle Port Intelligent Inspection System
Integrating UAV control, machine vision and digital twin technology, the team developed the intelligent inspection system of port intelligent identification and space-time twin monitoring UAV, and realized the full life cycle inspection of pre-warning, in-process analysis and post-event backtracking.
The team built a port safety management and operation early warning and monitoring system to realize real-time dynamic display of production equipment status and alarm information, real-time dynamic tracking of the whole operation process of dangerous goods, real-time monitoring of port safety production and operation, and intelligent judgment of potential risks.
Figure 10 Regional temperature anomaly warning(Sample)
The self-developed UAV control APP intelligently sets the optimal inspection path of the UAV, realizes the dynamic cruise management of the UAV, real-time video data return and historical inspection task backtracking, and solves the problems of high risk of personnel working at altitude and high cost of manual inspection.
Figure 11 Set the UAV inspection path(Sample)
Comprehensively collect and store port IoT real-time data and early warning information, integrate the twin platforms to accurately reproduce the historical operating environment and risk situation, identify bottlenecks and deficiencies in operations, so as to find potential problems and optimize future security management and control strategies.
·END·
上海海事大学港口物流数智化研究团队
上海海事大学港口物流数智化研究团队通过OR运筹优化、AI人工智能、VR虚拟现实、MV机器视觉、DT数字孪生等前瞻性技术的研发和应用,推进港航物流关键节点的数字化和智能化改造。
团队始终致力于港口智能化和数字化领域的“产学研用”融合发展,以实际行动推动港口产业数字化创新。
作者:海大港口数智团队;转自:咪啰科技
本公众号注明“来源:XXX”或“xxx提供”的作品,均转载自其他微信公众号媒体或网站,转载目的在于分享行业信息,不为其版权负责。如果您发现有不当之处或有侵犯您的知识产权的作品,请及时与我们联系,我们将进行修改或删除。