top of page
Vinemapper blanc.png
Sans titre - 6.png

On-board sensor for anomaly detection

The VineMapper is a tractor-mounted sensor capable of detecting vine anomalies and creating a map of affected vines in real time.  A small control screen installed in the cabin controls the device. Once switched on, it automatically detects when the tractor is entering or leaving a plot, and can identify the plot it is in. Once inside the plot, it will take photographs of the vine rows using a very high-resolution camera as well as a high-performance flash, at a rate of two images per second, enabling the tractor to travel at speeds of up to 7 km/h.

As the VineMapper is installed on the side of the tractor, photos are taken in the trellising plane. This enables it to spot a large number of anomalies, giving much better coverage of the vineyard than if it were mounted on a drone, for example.

After each photo is taken, our AI analyzes the image to detect anomalies in real time. This efficiency is made possible by the device's onboard computer: instead of having to send the images to a cloud where they will be analyzed remotely, which can take several hours, they are analyzed immediately after being taken. Then, the nature and position of the anomaly, accurate to the nearest centimeter thanks to an RTK (Real Time Kinematics) antenna integrated into the device, are sent to a database so that they can be visualized in the form of a map on VineCore.

Photo VineMapper.png
Cartographie VineMapper sur fond vert.png

The AI was trained on vineyard images taken by the VineMapper, which were then annotated on our own annotation platform, AgroMark. The particularity of this platform lies in its moderation system, which enables experts to control the quality of annotations, thus guaranteeing the quality of the data feeding the model, and hence its reliability.

VineMapper can be rented to anyone with a subscription to the Agronomy Monitoring.

bottom of page