Short description of the project
In the BMBF project AgriApps (application modules for the automated management of special agricultural crops) a modular concept for the automation of agricultural tasks is being researched. An autonomous robot can be equipped with different "apps" that perform different tasks. Thus, a high utilization of the robot over different seasons and thus an economical operation can be achieved.
As an example, the project developed an app for mechanical weed control in special crops. Several rotary harrows remove the weeds and loosen the soil while the robot drives autonomously through the crop. An actuator system allows the positioning of the mechanical manipulation unit so that weed control is possible both between and within plant rows.
Recognition of useful plants and weeds
A prerequisite for this is the automatic differentiation of crops and weeds developed at the Fraunhofer IOSB. While the robot is moving, a multi-sensor module records the row of plants to be processed. Based on shape, color and texture characteristics, the plants are distinguished from weeds and soil. The classification software learns this distinction from a small amount of training data and can therefore be easily adapted to different application scenarios. The recognition results are used in real-time to control the manipulator for weed control.