Design and validation of a real-time cassava planter seed quality monitoring system based on optical fiber sensors and rotary encoders.

Targeting the issues of seed leakage and cutting segment adhesion due to poor seed feeding and cutting in real-time seed-cutting cassava planters, this study developed a seeding quality monitoring system. Based on the structure and working principle of the seed cutting and discharging device, the installation methods of the matrix fiber optic sensor and rotary encoder were determined. By combining the operational characteristics of the planter's ground wheel drive with seed cutting and seed dropping, a monitoring model correlating the sowing parameters with seed dropping time was established; a monitoring window was created by extracting and processing the rotary encoder pulse signal, and the number of seeds sown after each opposing cutter's operation was calculated based on the pulse width information within the monitoring window. The monitoring system's statistics were compared and analyzed with the manual statistics, and the bench test showed that the monitoring system designed in this study offers high accuracy. When the simulated rotational speed of the opposing cutter ranges from 10 to 30 rpm, the average monitoring error between the monitored and actual seeding quantities for the left and right rows is less than 1.4%. The monitoring system can promptly and accurately activate sound and light alarms for faults, achieving a 100% success rate in alarms and an average fault response time of less than 0.4 seconds. Field tests demonstrate that the average error in seeding volume is 0.91%, and the monitoring system can timely alert to faults occurring in the planter. The system fulfills the requirements for real-time monitoring of cassava seeding volume at various operating speeds in field conditions, and can serve as a reference for monitoring operational parameters in subsequent cassava combine harvesters.