| At present,the agricultural industry is facing severe challenges,and the sharp increase in labor costs has become the main factor restricting the development of China’s orchard planting industry.Therefore,the autonomous operation of orchard machinery has become the main direction of the development of orchard machinery.In the actual environment of orchards,high-precision and stable positioning is the premise of independent mechanical operation of orchards,but due to the complex operating environment of orchards,the multipath effect formed by tree occlusion seriously affects the satellite positioning accuracy,therefore,on the basis of summarizing the relevant technologies of orchard positioning systems at home and abroad,this paper develops an orchard-assisted positioning mobile platform based on AprilTag visual reference system to assist GNSS-RTK signals to achieve high-precision positioning and improve positioning accuracy and stability.It provides a technical basis for high-precision positioning,autonomous operation and precision operation in orchards,and accelerates the development of intelligent agricultural equipment in orchards in China.The main work of this article is as follows:(1)A mobile platform for orchard-assisted positioning based on AprilTag visual reference system was built.The platform takes the crawler electric dual-motor drive chassis as the mobile platform,the Jetson AGX Xavier artificial intelligence processor as the computing platform,combined with ROS,AprilTag,depth camera calibration and other software systems,selects ZED2 binocular depth camera as the data acquisition end,GNSSRTK as the absolute position acquisition end of the reference point,obtains the relative position of the marker and the reference point through the AprilTag visual reference system,and obtains the absolute position of the marker through coordinate transformation.The test results show that the accuracy of the auxiliary positioning system is 99.8%,the recall rate is99.8%,and the main measurement error is basically controlled within 6cm.(2)The parameter index of the accuracy of the AprilTag visual benchmark system was constructed.Six sizes of QR codes of 5cm×5cm,10cm×10cm,20cm×20cm,30cm×30cm,40cm×40cm,50cm×50cm were selected for distance error analysis test.In an orchard environment,while larger sizes provide better robustness and reliability,larger size codewords have a higher minimum line-of-sight threshold.Through verification,the AprilTag of 19,133,272,426#30cm×30cm size is in the 300 cm range,and the AprilTag of40cm×40cm size is within 5cm in the 400 cm range.(3)An optimization scheme based on the positioning accuracy of orchards based on AprilTag visual reference system is proposed.The influence of different light intensities on the positioning accuracy of orchard-assisted positioning mobile platforms was explored,and the system had a better recognition effect during the noon period between 5000 and 10000 lumens.The orchard-assisted positioning mobile platform is visually processed,and the system after EKF smoothing is adopted,and the coordinate conversion is carried out after completion.After data comparison,the error threshold after optimization is reduced by33.33%,and the main measurement interval is reduced by 66.67%.The average measurement error of AprilTag in the size of 30cm×30cm,40cm×40cm is 5.8cm,the measurement error is maintained within 400 cm,the recognition accuracy reaches more than99%,the maximum response time of the system is 146 ms,the minimum value is 14 ms,the average value is 50 ms,and the typical response time of the system is 50 ms,which meets the actual needs in the orchard environment. |
PDF Full Text Request