Development of a stereovision sensing system for 3D crop row structure mapping and tractor guidance

This research aimed to develop a field sensing system capable of performing three dimensional (3D) field mapping for measuring crop height and volume and detecting crop rows in 3D for reliable tractor guidance using one tractor-mounted stereocamera. The core of the dual-application field sensing system is a stereovision-based mapping method for creating 3D crop structure maps by means of estimating the motion of tractor-mount stereocamera and stitching constituent stereoimages progressively. By this approach, a feature point tracking algorithm was used to extract tractor motion indicating feature points from constituent stereoimages, and then feed the outcomes to a mathematical model of tractor dynamics to estimate tractor travelling speed and heading direction. Field validation test results indicated that the developed system could accurately create a 3D crop row structure map that can truthfully represent the spatial variability of crop height and volume from acquired stereoimages and can be used for reliably guiding the tractor by following crop rows.