Design and Test of Cleaning Loss Kernel Recognition System for Corn Combine Harvester

Abstract

Cleaning loss is an important parameter to judge the performance of a corn combine harvester. At present, there exists the phenomenon that corn cleaning loss rate detection devices have a long signal processing time and low recognition accuracy. To solve this problem, based on the principle of the impacting piezoelectric effect, the impacting signals of corn kernels and impurities are analyzed by means of the frequency spectrum method to obtain the characteristic frequency for effectively distinguishing corn kernels and impurities, which is determined as 8.7 kHz. Based on this characteristic frequency, a corn cleaning loss kernel recognition system is designed, which can realize the function of corn kernel recognition and cleaning loss rate recording. In this system, signal processing circuits which mainly include two-order high-pass filtration, envelope wave detection and voltage comparison are designed. On the basis of the signal processing circuit, adding the judgment program for the holding time of the output square wave signal improves the system’s recognition accuracy for kernels impacting the sensitive plate simultaneously. The system was tested in indoor conditions. The results show that 20–30 corn kernels could be accurately recognized per minute on a single sensitive plate, and the recognition accuracy rate of this system could reach 85% when three corn kernels impacted simultaneously. The results serve as a theoretical basis and represent a new method for the design of a cleaning loss kernel recognition system.