BoniRob (Amazone-Werken, University of Applied Science Osnabrück, Company Robert Bosch)
BoniRob is an autonomus field robot promoted by the federal ministry for food, agriculture and consumer protection and the federal agency for food and agriculture in a joint project of the Amazone-Werken, the University of Applied Science Osnabrück and the company Robert Bosch. The field robot shall navigate for food and agriculture in a joint project of the Amazone-Werken, the University of Applied Science Osnabrück and the company Robert Bosch. The field robot shall navigate autonomously over field with the help of sensor data and preform tasks using different application modules. These modules are pulled in the module carrier BoniRob.
BoniRob Spraying Application (Cooperation University of Applied Science Osnabrück)
In modern agriculture ecological aspects and efficiency is becoming increasingly important. To meet these requirements the agricultural technology developes at high speed. Nowadays for example it is possible to adjust the output amount, the margin width and the throwing range of manure distributors. The University of Applied Science researches on further developments in the agricultural sector. They often work together with local businesses on solutions and improvements, as well as on new concepts. This master thesis was developed within the INTERREG, EUREGIO and the Productschap Akkerbouw “SmartBot” project and is working on such a concept. Core of the thesis is to develop a module that will enable a more efficient regulation of weeds compared to conventional practices.
The thesis deals with the development of an application module with camera-based selective weed control for an autonomous field robot. The application module is a system that detects weeds in corn rows by using consumer cameras and image processing. In this work the image processing was developed with the open source library “OpenCV”. The detection of the plants is based on color filtering in the HSV color space. Subsequently the objects have to be searched out of the filtered image to allocate the right spray zones of the nozzle arm. The overall result of this work is a robust application module for field robot “BoniRob, which works effective in detection of plants and in distinction of crop and weeds. Furthermore, the finished module can be used flexibly for other research purposes in order to test other camera systems or image processing algorithms.