Authors

Abstract

Most existing grain flow sensors are designed for paddle-type elevators, with limited focus on applications in auger elevators. This paper addresses the yield monitoring needs during rice harvesting operations, specifically targeting auger-based outlets through experimental research. An array-type differential grain flow sensor was developed and an indoor test bench was constructed to evaluate its performance. The study compares the effectiveness of time-domain and frequency-domain differential processing, alongside various filtering methods, for pre-processing the sensor’s raw signals. Additionally, a grain flow regression model was built using the Random Forest algorithm. Experimental results demonstrated that the monitoring errors during field tests ranged from -6.42% to 8.23%, indicating that the sensor met the requirements for rice yield monitoring. This sensor provides valuable data for feed rate detection, speed regulation, and adjustments to the threshing and cleaning systems in combine harvesters, offering significant practical implications for the promotion and development of precision agriculture.

Abstract in Chinese

现有的籽粒流量传感器大多面向于刮板式升运器，鲜有针对搅龙式升运器工作场景的籽粒流量传感器。本文面向搅龙式出粮口，针对水稻收获作业的产量监测需求进行试验研究，开发了阵列式指板差分结构的籽粒流量传感器，研制了籽粒流量传感器室内试验台，对比了时域、频域差分处理以及不同滤波方法对所采集的传感器原始信号的预处理效果，并基于随机森林算法构建了籽粒流量回归模型。试验结果表明田间试验监测误差为-6.42%~8.23%，能够满足水稻收获的产量监测需求，可以为联合收获机喂入量检测、前进速度调控或者脱粒清选装置的作业参数调节提供数据参考，对精准农业的推广和发展也具有重要的实际意义。