P. Reyns’s research while affiliated with KU Leuven and other places

During the summer season of 2000, a feasibility study is carried out to measure grain quality (moisture and protein content) with Near Infrared Reflection (NIR) technology on a conventional TX64 New Holland combine harvester. A Zeiss Corona 45 NIR 1.7 sensor was installed on a bypass of the clean grain elevator. Parallel with the NIR measurement, grain samples were taken at the end of the bubble-up auger that transports the grain from the clean grain elevator to the grain bin. Measured signals at the diode array of the spectrophotometer are highly variable in time and since data is acquired at a sampling rate of 23 Hz, an appropriate low pass filter is designed to obtain every second a relevant spectrum. In addition, an optimal time shift is calculated between the location of the NIR measurement and the sampling spot, in order to make the comparison between the estimated spectra and grain samples as accurate as possible. The spectra are first spectrally converted to either absorbance or Kubelka-Munk values. The required spectra of a white and black standard are time-interpolated between the standard measurements executed closest before and after the selected on-line spectrum. Preprocessing, calibration and validation are executed using the PLSplus/IQ module included in the GRAMS/32 software package. After application of Mean Centering, the Multiple Scatter Correction, Standard Normal Variate and detrending algorithm are applied. The calibration models are developed using the PLS algorithm and validated through cross-validation.

Nowadays, commercial combine harvesters are equipped with special sensors for the purposes of precision farming. Sensors are installed to measure the harvested surface and mass flow each second. Hence, grain yield maps are constructed, approximating the actual grain yield on the field before harvesting.Most grain flow sensors are installed on the paddle elevator before storage in the grain tank. At the elevator exit, paddles throw packs of grain against a curved plate on which a force transducer is attached. Therefore, the measured flow signal is heavily disturbed by the paddle rate and resonance frequency of the mechanical sensor.In this study, a computationally cheap method is presented to track both frequencies in the measured flow signal. Based on these frequency trackers, a double notch filter is tuned continuously to remove both disturbances. Moreover, the knowledge about the frequency content can be used to monitor the condition of the mechanical sensor and the paddle elevator.

An on-line crop density measurement delivers interesting information about the local crop condition. Once this measurement is established in a non-destructive way, crop density monitoring could be performed during the growing season and results would immediately be used for a site-specific application of nitrogen. In this work, some primary results are shown from an ultrasonic crop density measurement. A commercial sensor, normally used for level measurements, is equipped with an extra output, providing a signal corresponding with the ultrasonic wave transmitted through a band of crop. Because of the properties of the measured signal, a non-linear transformation of the measured signal is necessary. With the result of this static compensation, a good estimation can be found of the energy in the received wave, closely related with the volumetric density of the crop. Once the appropriate signal is found, an experimental set-up is constructed for dynamic crop density measurements. A small wagon is placed on two parallel rails with a stroke of crop in between. On both sides of this wagon, two ultrasonic transducers are placed. One as transmitter of the ultrasonic wave, the other one as receiver at the other side of the crop. With this arrangement, a first calibration curve is made, delivering an exponential relationship between the transmitted energy and volumetric crop density.

To maximize economic return from agricultural production units, costs have to be minimized and benefits maximized. For grain, kernel yield and quality have to be maximized while the use of seeds, fertilizer, herbicides and fungicides have to be optimized. The best location to evaluate productivity levels, by measuring yield and quality of grain and straw, is the combine harvester. Moreover, other grain quality characteristics like density or test weight can be determined for use as an evaluation tool. In this paper, an overview is given of the past and current research toward the evaluation of currently available commercial sensors (e.g., for measuring grain yield and grain moisture content) as well as toward the development of new sensors (e.g., grain protein content and straw yield).

The objective was to control technical factors and study the accuracy and precision of yield maps. First, an optimal data analysis and management method for raw data of precision agri-culture sensors was developed. Only then, we can start to analyse yield data. The use of a certain management and evaluation method of raw data has a crucial role, which was ne-glected in the past. Yield maps are made frequently, but the precision and therefore the use are very important to take into consideration. Because of many factors, the raw data of the yield mapping system contain a lot of unrealistic data, which most software packages cannot detect. The raw data have to be cleaned and made universally applicable and repeatable in or-der to allow comparisons of different data layers within a field and of differences between fields and crops.

Site specific information was acquired on a Belgian field during harvest in two consecutive years. Grain yield was recorded with previously developed sensors. At the same time, grain protein and moisture content were determined at different locations in the field. The first aim of this study was the mapping of grain moisture and protein content in order to determine the amount of variation within one field. Secondly, yield and quality maps were compared to each other and studied for relationships between the different parameters for different parts (growing conditions) of the field. Finally, these relationships were used to delineate different fertilisation zones in the field. A variation in moisture content of over 7% existed in the field in both years. Also for the protein content a variation of more then 4% was noticed. Variation in protein content varied from one year to another, possibly due to different weather conditions. Consequently, only 2% of the area of the field could be classified being over-fertilised.