Affiliations: Department of Physical Education, Yuncheng University, Yuncheng, China
Correspondence:
[*]
Corresponding author. Haiming Liu, Department of Physical Education, Yuncheng University, 044000, Yuncheng, China. E-mail: [email protected].
Abstract: In order to solve the problem that the seeder operated by hand controller lacks high-precision visual function, cannot self-identify the seed target, and has low sowing precision, this paper proposes a new automatic sowing robot control research based on basketball motion capture. Through motion capture, obtain the target characteristics. Based on the target characteristics acquired by the vision module of the robot, the path planning of the robot is realized by rotating a certain parking space Angle on the basis of real-time positioning. The performance test shows that the accurate recognition rate is higher than 95.2%, and the maximum is 98.1%. The accurate seed localization rate was 95.5% and the maximum was 98.2%. By comparing the trajectory precision of the joint space of the auto-seeding robot, the precision of the trajectory control method is better than that of the traditional method. The residual removal rate of this method is 39.87% – 46.25%. The effect of different placement methods on seed placement depth showed that the average depth value obtained by this method was 0.001.49, and the depth variation coefficient was 0.70% – 2.00%. This method can improve the target recognition rate and positioning accuracy, meet the needs of high precision sowing operations, and increase the crop yield.
