Contents lists available at ScienceDirect Computers and Electronics in Agriculture journal homepage: www.elsevier.com/locate/compag Original papers Assessing an infrared-based seed drill monitoring system under field operating conditions Hadi Karimi a , Hossein Navid a, ⁎ , Bahram Besharati a , Iraj Eskandari b a Department of Biosystems Engineering, Faculty of Agriculture, University of Tabriz, Iran b Agricultural Research, Education and Extension Organization, Dryland Agricultural Research Institute, Maragheh, Iran ARTICLE INFO Keywords: Seed drill Sowing performance monitoring Seed flow Field experiment ABSTRACT In this study based on the infrared seed sensor developed earlier, a seed drill monitoring system was designed and constructed. In the proposed monitoring system, seed flow sensors were installed individually on each seed tube so that the seed flow rate, as well as the presence or absence of seed flow, could be displayed on a graphical interface. The ground speed was measured with the Hall Effect sensor, and taking into account the mass flow rate of the seed, the sowing rate was calculated according to the seed mass sown per unit area. During operation, the system registered sowing data with the location information provided by the GPS module. The overall in- formation from the sowing performance was then recorded on the embedded memory card and simultaneously displayed on the graphical interface. The proposed system was also provided with ultrasonic sensors to con- tinuously measure and display the seed and fertilizer levels in the hoppers. The monitoring system was con- structed and installed on a seed drill equipped with 13 sowing units. With three levels of ground speed and sowing rate during field experiments, the sensing system was assessed under outdoor operating conditions. The field test resulted in a correlation coefficient of 85% between the mean of the weighted data obtained from the scale and the mass flow estimates. The obstruction of the optical elements by dust seems to have had the most adverse effect on the performance of the proposed sensing system. 1. Introduction One of the most important issues in growing crops is the use of a proper density of plants per unit area. The proper distribution of crops results in optimal utilization of environmental factors and lower levels of intraspecific competition in populated areas. During the course of their research, Kaya et al. (2007) and Sedghi et al. (2008) concluded that it is possible to achieve the highest level of yield by applying the optimum density for different varieties of crops while using optimum nutrients, moisture, and light. Sowing quality influences the effective- ness of weed control throughout the vegetation stage, as the chance of weeds spreading rises with an increase in empty spaces resulting from planting faults (Kostić et al., 2018). Overall, the optimal planting density per unit area increases the yield of agricultural products while preventing the overuse of such inputs as seeds, water, fertilizers, pes- ticides and herbicides (Anantachar et al., 2010; Yasir et al., 2012). Nowadays, producers would like to present planters with significant development in the precise distribution of seeds. The distribution of seeds per unit area across a field directly depends on the quality of the planter's performance during the planting operation (Maleki et al., 2006; Turan et al., 2015). Seed drills are built to place seeds into the soil at the desired density per unit area. In a typical seed drill operation, seeds are metered using fluted rollers, which release seed mass gradu- ally through a series of seed tubes. To optimize the distribution of seeds per unit area, it is common to calibrate the seed drills before sowing. Ordinary methods of calibration are generally tedious and time-con- suming. Owing to the nature of the fluted rollers, variation in sowing rates is considerable (Al-Mallahi and Kataoka, 2013; Maleki et al., 2006). Moreover, one or more of the seed tubes could become in- operative during field operation due to such issues as excessive or in- accurate sowing, delivery tube blockage, or emptying the seed hopper and damaging the transmission system. Therefore, although the amount of seed distribution is clear for decision makers, sowing rates may vary with the factors linked to the performance of the seed drill and the ground speed of the tractor during the sowing process (Quanwei et al., 2017). As the performance of the seed drill and its components is en- tirely invisible to the operator, no clues about sowing quality during field operation are available (Xia et al., 2010). Despite the presence of many factors affecting the quality of sowing, it is possible to achieve the desired plant population and increase yields if corresponding online https://doi.org/10.1016/j.compag.2019.04.045 Received 2 January 2019; Received in revised form 23 April 2019; Accepted 29 April 2019 ⁎ Corresponding author at: Department of Biosystems Engineering, Faculty of Agriculture, University of Tabriz, Tabriz, Iran. E-mail address: navid@tabrizu.ac.ir (H. Navid). Computers and Electronics in Agriculture 162 (2019) 543–551 0168-1699/ © 2019 Elsevier B.V. All rights reserved. T