Research On The Key Technology Of Vision-based Auto Navigation For Walking-type Rice Transplanter

As the growing of aging society problem, as well as rural labor transfer withindustrialization in China, intellectualization has become the inevitable trend of the futuredevelopment of agricultural machinery, autonomous navigation technology has become a hotissue of the international research in the field of agriculture machinery. Rice is one of the mostimportant food crops in China. Walking-type rice transplanter is widely used in southern Chinawith its compact, flexible, and energy-saving characteristics, but its labor-intensive, so carry outauto navigation research on the platform of Walking-type rice transplanter is very urgent.Compared to other navigation modes, the vision-based navigation system has the advantage ofeasy to build, great navigation accuracy and good real-time. Therefore, we try to find adistinctive way for transplanter automation, with combining the wireless remote monitoringwith vision-based navigation on the platform of walking-type rice transplanter. Main work ofthis paper can be summarized as follows:1) The design of assist walking mechanismDesigned and manufactured the assist walking mechanism with lifting function. On thebasis of analyzing transplanter’s Original traveling mechanism and lifting mechanism, thispaper designed two schemes, analyzed and calculated the linkage length of the first scheme.Experiment confirms that, through the assist walking mechanism, the transplanter has betterstability with three wheels touchdown when walks independently, which is beneficial toautonomous working; Furthermore, in order to remain the viewing angle of camera unchangedto capturing more valid video information used in navigation, it also makes sure front portionand rear portion of the transplanter ascend or descend equal height.2) The construction of the wireless remote monitoring systemThe paper constructed a wireless remote monitoring system based on the rice transplanter,accomplished developing of autonomous navigation system software of host computer based onVC++6.0, designed lower computer MCU_based control signal processing system andpneumatic actuator system, and instructed design method and concrete implementations ofsoftware and hardware of systems in detail. Experiment result show that: Real-time Imagetransmission, the control signal transmission delay in10ms of less than, pneumatic systemimplement motion control instruction timely, reliable, meeting the system functional requirements.3) The designing Vision-based navigation systemThe paper researched the suitable color model used in image processing for paddy fieldenvironment, introduced a camera calibration method based on OpenCV, studied the seedlingline extraction algorithm based on Hough transform, designed and implementation thenavigation parameter acquisition algorithm. Finally, input navigation parameters to the fuzzycontroller, and send the output motion control instruction to the lower machine, control thetransplanter tracking navigation path driving. The simulation results show that the controllercan well realize transplanter for navigation path tracking.