Orchard tree-shaped frame type autonomous reasoning and scene understanding method of robot

Abstract

The invention provides an orchard tree-shaped frame type autonomous reasoning and scene understanding method of a robot, and belongs to the field of agricultural robots. For a semi-structured orchardcomplex environment, the robot autonomously navigates and locks a crown through a laser radar to obtain main branch-column frame skeleton features of the crown, the features are compared with featuresin an orchard agronomic knowledge base built in a control system to realize autonomous reasoning of an orchard tree-shaped frame type, then a corresponding eye-in-hand camera initial positioning modeis automatically selected on the basis of different orchard tree-shaped frame types, crown three-dimensional contour coarse imaging is completed, and finally a corresponding crown partition mode is automatically selected to achieve close-range scene understanding, target recognition and positioning. According to the method, the problems of poor adaptability and extremely low efficiency of complete dependence on groping type detection of an unknown orchard scene during robot operation are solved, and autonomous continuous understanding of the orchard scene is realized.

Description

technical field [0001] The invention relates to the field of agricultural robots, in particular to a method for autonomous reasoning and scene understanding of a robot's orchard tree frame. Background technique [0002] The orchard is a typical semi-structured complex environment, and the effective understanding of the orchard scene is the core to ensure the autonomous continuous operation of the robot. Based on a full understanding of the orchard scene, the robot can navigate in the orchard and make decisions about complex actions such as continuous fruit picking, bagging, pruning, leaf threshing, and pollination. However, the existing research is still difficult to meet the needs of continuous operation of the robot in the orchard site, the main problems are as follows: [0003] (1) There are great differences in orchard construction specifications and tree frame types. Existing navigation and operations either rely on prior information and pre-navigation mapping, or comp...

AI Technical Summary

Problems solved by technology

However, the existing research is still difficult to meet the needs of continuous operation of the robot in the orchard site, the main problems are as follows:

[0003] (1) There are great differences in orchard construction specifications and tree frame types. Existing navigation and operations either rely on prior information and pre-navigation mapping, or completely rely on the robot's exploratory detection of completely unknown and complex orchard and tree crown environments , the huge computing power consumption leads to poor adaptability and extremely low efficiency of the robot;

[0004] (2) Existing research mainly focuses on single research on tree row detection and fruit segmentation and extraction, but lacks the autonomous and continuous understanding of different scales and different tasks in the continuous operation process of robots in orchards, which cannot meet the needs of robots walking in orchards, Complex decision-making needs for finding goals and jobs

Images