Development and Evaluation of a Guidance-Assistant Algorithm for Agricultural Tractors in Plowing Operation

Abstract

This paper describes a robust procedure to obtain a guidance directrix for agricultural tractors in plowing operation with the wide toolbar. The
automatic steering algorithm controls the vehicles based on the position of the trajectory results of previous operation on the field. Hough transform
method was adapted to process forward view images of a vehicle captured by a digital camera installed at 1.1m height and angle of 35o respect to
line of travel. The contrast between disturbed soil and undisturbed soil was the basis for detection of the last track of the machine. This algorithm
was tested in the various situations including various illumination conditions and various plant residues. The position of the detected line respect
to the furrow was measured by algorithm. The methodology devised to overcome the image noise problems and to be able to determine the proper
trajectory for the tractor. The algorithm was able to find the trajectory with high accuracy. Maximum deviation of the detected line from the desired
trajectory was measured up to 1.72cm. The suggested algorithm is able to assist the operator during manual guidance of wide plowing tools.