Feasibility Study of Using a System Based on Image Processing for Guidance of Tractors During Plowing Operations

 Nowadays, tractors are almost driven by drivers. A driver does not only has to steer it, but also keep the mounted implements in a given position for some operations, including soil tilling, seed planting, weed controlling, spraying, … . In tillage operation, tractor and attached implements guidance is obtained by fallowing a furrow created by the plow in the soil during the previous passage. Driver has to fallow it to ensure a correct lateral position, but this continuous tiresome work affects his performance. In this research, a tractor driver assistance system was implemented for tillage operation. This system design was based on image processing technique. A test frame was designed and fabricated and a camera and lightening sources were mounted on it. The different positions of the camera on the frame were evaluated to find the best installation place. The experiments were conducted on some different farms. The important factors that affected the system accuracy including soil moisture content, environmental conditions and sunny or cloudy day situations were evaluated. The effects of farm surface conditions, last crop residues, natural or artificial light sources on the system performance were evaluated. LED lightening sources showed the best result, especially when they were used horizontally. Artificial lightening did not show visible differences with natural lightening in dry farming and sandy soils. Remaining cereal stubbles caused the worse system inaccuracy between the studied farms.