Remnant fertilizer monitoring system for maize fertilizer applicators

A remnant fertilizer monitoring system utilizing three-dimensional (3D) reconstruction was proposed to detect the amount of remaining fertilizer in the applicator's tank. Bench tests were carried out to compare the performance of four algorithms to estimate the remnant fertilizer amount: fertilizer remnant monitoring biharmonic spline algorithm (V4), natural nearest-neighbor algorithm (Natural), linear algorithm (Linear), cubic algorithm (Cubic). The average relative error for remnant fertilizer monitoring is 7.33% for the Linear algorithm, 7.30% for the Natural algorithm, 5.18% for the Cubic algorithm, and 4.30% for the V4 algorithm. Field tests are conducted at three fertilization rates to compare the performances of the V4 and Cubic algorithms. The average relative error for discharged fertilizer monitoring is 8.64% for the Cubic algorithm, which is 1.91% lower than that of the V4 algorithm. The results show that the Cubic algorithm has the best performance for remnant fertilizer monitoring. The average relative error of remnant fertilizer monitoring is 2.42% for the Cubic algorithm, which is 0.43% lower than that of the V4 algorithm. The response time of the remnant fertilizer monitoring system is 0.26 s. The results demonstrate that the proposed remnant fertilizer monitoring system is highly accurate and suitable for real-time applications.