RESUMO
In order to obtain desirable crop yields, grain seeds need to be sown at the optimal seed amount per hectare with uniform distribution in the field. In previous grain sowing processes, the seeding rates are controlled by the rotational speed of the flute roller which significantly effects the uniform distribution of the seeds due to disturbances, such as the reduction of the seeds' mass in the hopper and the change of working length of the flute roller. In order to overcome the above problem, we developed an adaptive roller speed control system based on the seed flow rate sensor. The developed system can monitor and feedback actual seeding rates. In addition, based on the monitoring value of the real-time seeding rates, an adaptive roller speed control method (ARSCM), which contains an algorithm for calculating the seeding rate with a compensation, was proposed. Besides, the seeding performance of the ARSCM and that of the conventional roller speed control method (CRSCM) were compared. The results of constant-velocity experiments demonstrated that the accuracy (SA) and the coefficient of variation (SCV) of the seeding rates controlled by the ARSCM were 94.12% and 6.77%, respectively. As for the CRSCM, the SA and SCV were 89.00% and 8.95%, respectively. Under variable-velocity conditions, the SA and SCV of the proposed system were 91.58% and 11.08%, respectively, while those of the CRSCM were 88.48% and 13.08%, respectively. Based on the above results, this study concluded that the ARSCM is able to replace the CRSCM in practical sowing processes for the optimal and uniform seed distribution in the field.
RESUMO
Having the correct seeding rate for a unit area is vital to crop yields. In order to assess the desirable seeding rate, the number of discharged seeds needs to be monitored in real-time. However, for small seeds, the miscounting of seeds during monitoring happens frequently when using conventional seeding quantity sensors, which have wide light beam intervals. Thus, a seeding quantity sensor, which enables small seeds to pass through the light beam steadily, was developed. Based on the seed-shading time, a seed-counting algorithm was proposed. Moreover, the key structure parameters of the proposed sensor were ascertained using an optimization experiment. Finally, the developed seeding quantity sensor was tested against a photoelectric sensor and a fiber sensor to compare the seed monitoring accuracies. The results show that the average monitoring accuracy of the developed sensor, photoelectric sensor, and fiber sensor were 97.09%, 56.79%, and 91.10%, respectively. Furthermore, the factorial analysis shows that the forward velocity of the experimental apparatus and the rotational speed of the seeding plate did not significantly change the monitoring accuracies obtained by the developed sensor. Therefore, the developed sensor can be applied to monitor the seed quantity for the precision seeding of small seeds accurately and robustly.