Using bi-spectral imaging technology for simulated online-weed control in winter wheat and maize

In spring 2011, two field trials on site-specific weed control in winter wheat and maize were carried out at Ihinger Hof research station of the University of Hohenheim. For the image acquisition, bi-spectral cameras were mounted on a vehicle. These cameras are able to take images free from disturbances by soil, mulch and stones. Images and the corresponding GPS-data were stored on-the-go. Afterwards, the images were analyzed by a weed recognition software. Weed infestation was mapped in consideration of weed species and weeds grouped according to their herbicide sensitivity. In order to simulate an online herbicide application, a onesided moving average of order five was used for the weed mapping. This kind of rearward calculation uses only the data of weed infestation which were already assessed behind or directly in the current position of the vehicle. The calculated weed distribution maps were checked by visual grid sampling. Herbicide application maps were generated by applying weed thresholds on the weed distribution maps. The herbicide application based on the maps was conducted by a multiple sprayer which allows the application of up to three herbicides independently from each other in a single pass across the field. Later on, the performance of the herbicide application was controlled again by visual grid sampling. Compared to a uniform herbicide application, the sitespecific weed control saved 83 % and 58 % herbicides respectively in winter wheat and 66 % in maize. The average efficacy of the site-specific herbicide application system in winter wheat was 70 % of the conventional herbicide application.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.