picture about gvmc and cars
AI Cloud

GVMC: Instant vehicle monitoring

Thousands of identical or similar vehicles are moving in the same logistics park. How can we distinguish in a short time the one we need information on? And this is in the context of high vehicle turnover: during 2018 and 2019, more than 10 million units were stored or transported in Europe’s ten largest vehicle logistics ports. Some of these centers could have between 750,000 and one million at any time. In addition, some centers want to increase the facilities by building between 10,000 and 15,000 square meters more.

The GEO-VISUAL-MULTI-CAMERA (GVMC) project seeks to position logistics vehicles using cameras in ports where thousands of similar cars move around. The multi-camera system developed for this purpose will precisely arrange each car’s location in real-time.

The consortium consists of Plain Concepts in an artificial intelligence technology leadership role and IberAmbar, a Portuguese logistics and communications hardware infrastructure business expert.

  • Full project title: GEO-VISUAL MULTI-CAMERA
  • Project file code: GA 864266
  • Project start and end date: 12/01/2021 – 11/30/2021
  • Budget: 109,500.65 euros
  • Project funding agencies: European Union’s Horizon 2020 research and innovation program and DigiFed
european union
digifed logo
Client
GVMC
Industry
Logistics Research
Services
AI / Cloud
Technology
Azure, Computer Vision, Deep Learning
This project has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement no. 872088 (GEO-VISUAL MULTI-CAMERA)
picture about gvmc, deep learning and computer vision
01

Challenges

In recent years, European automotive logistics parks have seen the number of vehicles transported or stored increase to more than 10 million per year, driving insurance costs to more than $7.7 billion. Monitoring each car is time-consuming and costly.
By adding Computer Vision, AI, Deep Learning, and geopositioning to surveillance systems and collecting and analyzing relevant data, customized search, and alert criteria can be created to locate vehicles instantly.
To train the technology, a data set consisting of 85 minutes of 1080-pixel, 60-frame-per-second videos were used. These videos were recorded by eight synchronized static and outdoor cameras.
picture about gvmc and cameras
02

Results

The project’s main objective, which is tracking identical vehicles using computer vision without the need for tagging, has been achieved.
This camera system can be added to the existing camera system in each port and could be adapted to other types of objects, such as cargo containers.
picture about geo visual multi-camera
03

Highlights

GEO-VISUAL MULTI-CAMERA seeks to reduce the problem of finding cars with a solution that reduces costs and operates in real-time, 24 hours a day, seven days a week.

Related Projects