DYNAMIC DECISION-MODEL FOR WEED-CONTROL METHODS IN CEREALS BY MEANS OF DIGITAL IMAGE-ANALYSIS

Field experiments were conducted to study weed-crop competition in winter wheat, -barley and -rye. Weed cover and crop cover was determined over time by digital image analysis of slides taken in the field. The relation of percentage weed coverage and crop coverage was most suitable to quantify the competition of weed population on the crop. In addition the model ''UNPROG'' was established to predict yield losses caused by weed competition and to estimate the profitability of chemical or mechanical weed control. The model was tested on 22 farms in Nordrhein-Westfalia in autumn 1991 and spring 1992. In 65 out of 72 field plots (92 %) of winter cereals ''UNPROG'' gave correct prediction with respect to necessity of weed control.