Path recognition method for vision navigation of tobacco harvester

Tobacco harvesters are used in large-scale reclamation area. The demand for automatic navigation is urgent because of long time and intensive labor. According to the specificity of tobacco fields' environment, the conventional path recognition method had been improved, which was taking 2G-R-B processing, threshold processing in turn and adopting dilation algorithm in image morphology instead of median blur algorithm. At last, modified Hough transform was used to extract navigation line. Processing results show that the algorithms which can be underlain in vision navigation research of tobacco machinery can meet the demand of vision navigation in tobacco fields and obtain the target navigation path rapidly and reliably.