| Ningxia is one of the best producing areas for wine grapes.Compared with the continuous expansion of Ningxia wine grape planting scale,the intelligentization of grape planting equipment is still in its infancy,which restricts the development of the grape industry to a certain extent.In order to better realize the automatic navigation of vineyard operations,an automatic navigation system based on lidar and visual research is suitable for operations in the vineyard environment.The navigation system perceives environmental information and makes navigation control decisions,combined with navigation system motion actuators to realize navigation motion control,and correct lateral deviations caused by the environment.The main research contents are as follows:(1)Navigation system hardware selection.According to the navigation operation requirements of Ningxia wine grapes,the overall scheme design of the navigation system is proposed.The navigation system is divided into environment perception layer,decision control layer,and motion execution layer,and the navigation system suitable for vineyard operations is studied around each layer.According to the vineyard operating environment and navigation requirements,the hardware of each part of the navigation system is selected,and then the overall construction of the navigation system hardware platform is realized.(2)Research on path planning of vision-based navigation system.Collect road images of vineyards in different seasons and different weathers to construct a semantic segmentation dataset of vineyard roads.Based on the research of the environmental classification elements of the vineyard road,it can judge the type of the grape area and the road area after the input,and build the vineyard road semantic segmentation model.Perform edge detection on the left and right boundary lines of the segmented road,divide the region of interest,fit the left and right boundary lines of the road through Hough transform and least squares method,extract the midpoints of the left and right boundary lines to determine the navigation line,and determine the navigation steering based on the mobile platform’s own pose Angle,realize the research of path planning based on vision navigation system.(3)Research on path planning of navigation system based on lidar.The lidar scans the vineyard road information to generate a vineyard submap,loop detection reduces accumulated errors,and generates a vineyard map through global optimization.The lidar continuously scans and updates the information,corrects its own pose,and realizes the mapping and positioning of the navigation system.Navigation path line detection is divided into two parts:global path planning and local path planning.Global path planning introduces graph search,and uses Dijkstra algorithm to plan the navigation line to the end point.Local path planning follows the global path through TEB algorithm and avoids it in real time.Obstacles,to realize the research on the path planning of the navigation system based on lidar.(4)Research on control method of navigation system.The mobile platform adopts the differential motion mode of left and right track wheels,analyzes the motion mode of the mobile platform,and establishes the kinematics model of the navigation system.The system steering control adopts the proportional-differential-integral control method,which realizes the steering motion control of the mobile platform by controlling the relationship between the speed of the left and right track wheels.Analyze the movement state of the mobile platform,and verify the three states of curve movement,linear movement,and rotation movement on this basis.Control the decision-making information of the navigation system and execute the navigation movement to realize the research of the control method of the navigation system.(5)Test results and analysis.Through the vineyard navigation test,the feasibility of the performance of the mobile platform,the accuracy of the navigation sensor and the performance of the navigation system was verified according to the test purpose and operation indicators.The test results show that when the mobile platform navigates at speeds of 0.5 m/s and 1.0 m/s,the average lateral deviation based on the visual navigation test is 3.3 cm and 3.5 cm,the maximum deviation is 7.3 cm and 9.7 cm,and the mean square error is 2.9 cm and 8.0 cm,the average lateral deviations of the lidar-based navigation test are 3.4 cm and 3.6 cm,the maximum deviations are 9.1 cm and 9.8 cm,and the mean square deviations are 4.2 cm and 4.2 cm.The test results show that the automatic navigation system basically meets the requirements of automatic navigation operations in vineyards.The mobile platform performance,navigation sensor accuracy and navigation system performance of the navigation system meet the requirements of navigation operations,which is feasible. |
PDF Full Text Request