Evaluation of yield sensors for site-specific management

Abstract

The performance of an impact-based yield monitor was investigated via field and laboratory experiments. In another study. X-ray techniques were used to examine the potential use of x-rays for yield measurements of harvested crops. Field and laboratory experiments showed that proper yield sensor calibration is die most important factor to keep the error less than± 4% when individual grain tank loads are considered. Field and laboratory tests provided evidence of decreasing error with increasing harvest area. In field tests, the accuracy of yield monitors increased consistently when yield measurements were averaged over IS, 30, and 60 m long segments, respectively. In laboratory experiments, the fluctuations observed in instantaneous yield data suggested that yield signals should be averaged over 4 to 10 seconds to attain the accuracy found for individual grain tank loads. Therefore, the accuracy determined at the end of harvest strips, using a weigh wagon as a comparison reference, should not be considered as true accuracy for small area applications. Grain yield variations can have an impact on accuracy. The smallest error occurred at constant flow rates whereas yield monitor error increased when step changes and transient flow rates were induced using the laboratory test stand. Another factor affecting the accuracy of the yield monitor was the combine ground speed. Changing the ground speed, depending on yield variations, also increased the yield monitor error. The accuracy of a yield monitor then would change across a field due to spatial yield variability and changing combine ground speed. The yield sensor's response to sudden flow rate changes was quite satisfactory. X-ray techniques provided high accuracy in com flow rate measurements showing repeatable correlation coefHcients of 0.99 between gray scale values and mass flow rate ranging from 2 to 6 kg/s. It is possible to increase the dynamic range of flow rates by making proper arrangements in the test setup and using different algorithms.