Machine Vision Based Automated Tractor Guidance

Driving an agricultural tractor in the field to perform various farming operations is typically hard for operators due to the long hours of tedious repetitive tasks. Among those tasks, guiding the tractor along crop rows without overrunning plants requires the most attention and is the major cause for operators' fatigue. A smart tractor guidance system, based on artificial vision using a CCD camera, can release the human operator from the monotonous driving task of following crop rows precisely. Not only will it release operators from the tiresome crop row following task but it will also allow them to pay more attention to other operations which will lead to improved performance, increased efficiency, and enhanced safety in farming operations. This paper presents the development of a machine vision based smart navigation system to guide an agricultural tractor following crop rows automatically and accurately. A video camera was installed on the front side of the tractor, and the developed algorithm processed the acquired images and generated appropriate steering signals to steer the tractor. Field tests verified that this smart guidance system could accurately guide a tractor along crop rows in fields at a traveling speed up to 11 km/h.