Tomato picking method and system of binocular robot based on YOLOv4 algorithm

Abstract

The invention discloses a tomato picking method and system of a binocular robot based on a YOLOv4 algorithm, and relates to the technical field of intelligent crop picking. The method comprises the following steps: establishing a sample data set of a plurality of occlusion degree categories; lightweight processing is carried out on the YOLOv4 algorithm, model training is carried out on the sample data set, and a target detection model is constructed; the parameters of the binocular camera are associated with the YOLOv4 algorithm; collecting and importing real-time tomato picking data into the target detection model through a binocular camera of the binocular robot, and generating tomato identification and positioning information; position information of a mechanical arm of the binocular robot is obtained, and grabbing track planning information is generated; and controlling the binocular robot to grab the target tomato. According to the method, lightweight processing is carried out on the YOLOv4 algorithm, a binocular vision technology is combined, and mature tomato fruits with different shielding degrees are divided, identified and positioned, so that the picking accuracy is ensured.

Description

technical field [0001] The invention relates to the technical field of crop intelligent picking, in particular to a tomato picking method and system based on a YOLOv4 algorithm with a binocular robot. Background technique [0002] At present, the development of modern agriculture is rapid, and the tomato arable land infrastructure construction area and total production volume are in a state of steady increase. The demand for automated agricultural smart equipment is also rapidly increasing. However, in the process of tomato production and processing, tomato fruit picking is a particularly labor-intensive and time-consuming step. The tomato picking machine came into being, it can work for people, can effectively reduce the intensity of labor and improve the production efficiency in the labor process. [0003] In the environment of smart agriculture, picking robots are developing more and more in the direction of intelligence, but there are still some deficiencies in refined...

AI Technical Summary

Problems solved by technology

[0004] (1) In the application of deep learning to the identification of ripe tomato fruits, the actual operation requirements of picking robots are not considered. Due to the limitation of the robot's viewing angle, there will inevitably be a large number of fruits blocked by tomato branches and leaves in the camera, but the current recognition is only to identify Tomatoes do not divide the degree of occlusion. As a result, the tomato picking robot cannot plan the trajectory of the robotic arm well during the process of picking mature fruits that are occluded, and there are shortcomings such as damage to tomato plants, fruits, and damage to the robot, resulting in economic losses.

[0005] (2) Due to the limited performance of edge devices, their computing power cannot directly undertake the YOLOv4 algorithm, resulting in low accuracy of device operations

[0006] (3) At present, most of the algorithms only complete the recognition of tomato fruits, but do not locate the recognized fruits, resulting in inaccurate target capture

Images