Pitaya pose estimation and stem shear point positioning method based on three-dimensional key point detection model

Through the method based on the three-dimensional key point detection model, the positioning and shearing points of the dragon fruit fruit are accurately estimated, which solves the problem of precise grasping and shearing of the dragon fruit picking robot, and realizes the lossless picking and preservation of the dragon fruit.

CN119722801AActive Publication Date: 2025-03-28GUANGXI UNIV
4 Cites 0 Cited by

Patent Information

Application Number
CN202411909151.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-24
Publication Date
2025-03-28
Estimated Expiration
2044-12-24

AI Technical Summary

Technical Problem

Existing dragon fruit picking robots are difficult to accurately grasp and cut dragon fruit, resulting in easy destruction of the peel and affecting the preservation and transportation of the fruit.

Method used

Using a method based on the three-dimensional key point detection model, the three-dimensional coordinates of the dragon fruit are obtained through the dragon fruit RGB-D image data acquisition system, a three-dimensional key point detection model is constructed, and the three-dimensional coordinates of the pose points of the dragon fruit fruit and the shear points of the corresponding branches are predicted.

Benefits of technology

It realizes the accurate estimation of the pose of the dragon fruit and the precise positioning of the branches shear points, solves the precise grasping and shearing problems during automatic picking of dragon fruit, and ensures the lossless picking and freshness of the fruit.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a pitaya pose estimation and fruit stem shear point positioning method based on a three-dimensional key point detection model, and the method comprises the steps: collecting the RGB-D image data of pitaya, constructing a point cloud set of pitaya fruits and branches, carrying out the pose fitting, obtaining the spatial three-dimensional key point coordinates, and carrying out the positioning of the three-dimensional key point coordinates. Segmenting and marking the pitaya fruits and branches of the color image, constructing a Mask R-CNN-based three-dimensional key point detection model, and finally training the three-dimensional key point detection model by using segmentation and marking data and three-dimensional key point coordinate data. And predicting a posture vector when the pitaya is grabbed and a spatial three-dimensional coordinate of a stem shearing point when the pitaya is sheared, and realizing posture estimation and three-dimensional stem shearing point positioning when the pitaya is grabbed and sheared.
Need to check novelty before this filing date? Find Prior Art