基于时变窗口的油菜精量排种器漏播实时检测方法

为了快速准确地反映不同漏播状态类型,针对油菜精量排种器漏播实时检测问题,在界定了"稀疏缺苗"和"断条"2种不同漏播类型的基础上定义了稀疏缺苗系数与断条系数,分析了两系数的二维平面分布,提出了一种采用排种盘转速与排种脉冲同步检测、排种频率与时间间隔双重阈值约束的检测方法。该检测方法利用MSP430单片机的时间捕捉中断资源,在由排种盘转速脉冲序列信号确定的时间窗口内,通过光纤传感器感应获得排种序列的排种频率与排种时间间隔2类统计指标,依据设定的漏播严重程度评判系数,建立了漏播状态评判规则。在气力式油菜精量排种器上试验结果表明:该检测方法能够根据排种盘转速的变化自行调整检测时间窗口,对"稀疏缺苗"和"断条"漏播状态判断准确率可达100%,对于油菜精量联合直播机在田间不同作业速度下,该方法能够有效实现精量排种器漏播的实时检测,在考虑地轮滑移情况下检测结果滞后距离低于(84*型孔数)mm。该方法不仅有效解决了小粒径精量排种器重播条件下漏播的检测,而且对播量不足进行了约束,提高了漏播检测的实时性,为后期实现精准自动补种创造了有利条件。

Method of real-time loss sowing detection for rapeseed precision metering device based on time changed window

Abstract: Loss sowing is one of the important indicators of measuring metering device quality and performance. Seeding faults such as the lack of seeds in the seed box, clogged seeding holes, transmission fault, and faster rotation speed of seeding plates will lead to loss sowing during the seeding operation. Numbers of continuous clogged seeding holes will cause periodic blank misses in a row and lots of discontinuous clogged seeding holes will produce periodic sparse misses. Both misses affect crop yield seriously. In recent years, lots of miss detection methods such as the time-interval method and seeding-frequency method have been studied on the precision planting to improve seeding performance, but few studies focus on the loss sowing detection in real time, especially for small seeds such as rapeseeds. In this study, pointing to the problem of how to detect the misses of small seed precision planting in real time, based on the concepts of sparse miss coefficient and blank miss coefficient, and the relation between two coefficients of two-dimensional plane and loss sowing, one double-threshold (including seeding frequency and seeding time interval) decision method has been put forward, in which a standard seeding pulse sequence has been developed to sense the seeding speed. The detection system utilizes time capture interrupts of MSP430 to acquire the seeding signals including seeding frequency and seeding time intervals by a fiber optic sensor within the time window determined by the finite pulse sequence of seeding speed. Based on the sequences of seeding frequency and seeding time intervals, two statistic parameters have been counted. With one setting coefficient which indicates the degree of miss, the two parameters can be used to decide the miss type. Tests were carried out on the JPS-12 test-bed of the pneumatic rapeseed precision metering device. During the platform tests the negative pressure was set to 900 Pa, the positive pressure was set to 500 Pa, and the speed of seed conveyor belt was 1.9 km/h. The loss sowing detection effect of pneumatic rapeseed precision metering device with 20 holes and 40 holes had been tested at the different rotating speed such as 12, 15, 18, 21, 24, 27, 30 r/min. One threshold coefficient of loss sowing was set to decide whether the metering device had produced miss and which type the miss was. During one time window, which was defined as the time of the seeding plate rotating a definite number of rounds, the detection system calculated the sparse miss coefficient and blank miss coefficient from the information of seeding frequencies and seeding intervals. Additionally, the field test was carried on precision the combine planter of 2BYM6/8 for both rapeseed and wheat seed driven by a DFH-LX954 tractor. During the test, the planter was set at three different level of working speeds, and each speed lasted for 8 m. The tests of the pneumatic rapeseed precision metering device showed that: it is effective and necessary to establish two coefficients (sparse miss coefficient and blank miss coefficient) to detect sparse miss and blank miss especially under both situations of multiple seeding and inadequate seeding. The accuracy of the loss sowing detection was 100%. Meanwhile, this method can accurately detect the loss sowing of the rapeseed precision planter at different working speeds in the field. The miss detection method can adapt the time window for the variation of seeding speed and shows a good real-time detection performance. Concerning the slip of copying land wheel, the standard distance delay of miss detection is less than 84 k mm, which makes it possible to re-seed automatically in the miss place at the right time.