As yield data are collected automatically it allows for errors to be introduced into the data collected. Filtering processes might be able to eliminate part of the errors and improve the quality of yield data. The development of a filtering process may be set up based on the individual characteristics of each error. These errors may be related to incorrect position, too high or too low yield, swath width, grain moisture, null distance between points, filling interval and outliers. The filtering algorithm was implemented on six sets of data generated from six different commercial yield monitors. A geostatistical analysis of the data sets showed that the application of the filtering process decreased the nugget effect and improved the cross validation quality index.
precision agriculture; yield monitor; algorithm
