Combination seed drills are usually used to plant grain seed in early fall. Because of limited time, there is a growing demand for using high speed operating machinery such as combination seed drills working continuously during whole day. On the other hand, for soil preparation, there is the problem of soil disturbance and misshaped fields after potato or sugar beet harvesting operations. So, the trace of the marker cannot clearly be identified by the driver, especially at the nights. In the developed countries, Global Positioning System (GPS) is usually used to accurately detect the routes. However, it cannot be implemented in Iran due to the lack of enough accuracy. In this study, a Tractor Driver Assistance System (TDAS) was designed, developed and evaluated based on the machine vision technology. This system is able to create a new path with specific distance with respect to the previous planted path by identifying the residual line created by combination seed drills. Therefore, the driver can easily follow the correct path illustrated in the monitor. According to different intensity of light during day and night, some simulations were first carried out and a computer program was written. In order to set up the system, a 3140 John Deere tractor, a camera with wireless capability, a personal computer as the main processor and a container and chalk powder as the marker of the previously planted path were used. For evaluating the different parts of the system, a fully randomized block design was used with 20 replications for each treatment. The results showed that there was no significant difference at 5% level for different light intensities during the day, but there was a significant difference between treatments under the light intensities of the day and those tested at the night. However, the error in the system during the night was acceptable and less than ±4 in detecting the previously planted path. Moreover, increasing the transverse distance and speed resulted in no significant difference between the tested treatments in finding the previously planted path at probability level of 5%. However, with increasing the speed of the tractor, the driver error in pursuing the new planting path was increased. Keywords : Precision agriculture, Image processing, Autonomous machines, Hough transform.
