Manipulation Detection for Expensive Agricultural Research

Bayer breeds, monitors, and researches newly developed resistant cereal varieties and crop protection products in breeding and crop protection fields around the world. The cultivated areas are often leased, freely accessible plantations and fields. At these locations, there is always a risk of unauthorized persons trying to manipulate the research projects. For instance, they may sprinkle or plant foreign seeds or chemicals in the soil or simply destroy crops and fields.

Since research projects on crop protection and seeds usually run for many years and the manipulation is not always discovered immediately, such incidents can cause major setbacks and financial losses for the company. Bayer therefore has a strong interest in early warning of any unauthorized, suspicious activities in its high-risk fields so that they can be investigated promptly.

In Securitas, Bayer has a well-known security service provider at its side in addition to its own corporate security staff. Securitas has already been entrusted with guarding various properties of the Bayer Group, which it carries out from »Security Operation Centers« (SOC). But 24/7 monitoring of the geographically dispersed, often remote and constantly changing test field locations is not yet possible.

»SensFArM«, short for »Sensor-Based Flexible Area Monitoring«, is a project in which the Fraunhofer FKIE has developed a deployable technical solution for sensor-based facilities monitoring. It is designed to ensure the detection and traceability of unwanted activity within the clearly defined areas of the Bayer test fields.

The technical approach is based on a multi-sensor motion detection solution. Seismic sensors detect vibrations in the ground. Additional optical detection and verification of movements is done by electro-optical and infrared cameras. Valid movement patterns are identified using sensor data fusion.

GPS own-position reporting via smartphones assists in differentiating between authorized and unauthorized access. The merged sensor data is transmitted via LTE from the field to the server. Alarms generated on the server are displayed on both stationary and mobile devices in a clear situational picture.

On the server side, rule-based processing of the merged sensor data is performed. An alarm is triggered in the event of a critical detection. Stationary and mobile applications offer an overview of the monitored areas as well as current alarms.

Sensor information and detected motion paths are georeferenced in a cartographically rendered layer. Alarm processing is supported in the application by enabling the user to classify and evaluate the detected objects, including via a video image.

A technical monitoring solution that meets the specific requirements is not yet available on the market. The »SensFArM« demonstrator can therefore serve as a jumping-off point for a fully mature technical solution that can be rolled out widely and offered by the industry. The generic approach to implementation pursued in the development of the demonstrator sets favorable conditions for further development stages oriented to specific applications.