Integratied Study of a Picking Robot for Greenhouse Grown Sweet Pepper

碩士 === 國立宜蘭大學 === 生物機電工程學系碩士班 === 105 === The objective of this study was to design a sweet pepper picking robot system, which consists of five major units: an end-effector, a machine vision, a robot arm, a vehicle and a central control. LabVIEW, a graphical programming language, was employed to dev...

Full description

Bibliographic Details
Main Authors: CHUNG,LI-WEN, 鐘立雯
Other Authors: CHIU,YI-CHICH
Format: Others
Language:zh-TW
Published: 2017
Online Access:http://ndltd.ncl.edu.tw/handle/7gw729
Description
Summary:碩士 === 國立宜蘭大學 === 生物機電工程學系碩士班 === 105 === The objective of this study was to design a sweet pepper picking robot system, which consists of five major units: an end-effector, a machine vision, a robot arm, a vehicle and a central control. LabVIEW, a graphical programming language, was employed to develop the system. The vehicle was modified by a four-wheel electric hydraulic lift platform truck, which can be used for autonomous walking between the inner walkway of the facility and equipped with a commercially available a six-axis machine arm, a machine vision and an end-effector. Machine vision used two color cameras to capture the image, and then program the image processing to determine the sweet pepper coordinates. Through a field investigation, the average diameter of a fruit stem was determined to be 11.7±1.8mm while the average weight of each fruit was found to be 118.7±17.79g. Use this information to design the end-effector with two-fingers, can be successfully applied to harvesting sweet pepper. The experimental results showed that the success rate of picking was 90% when the rate of fruit stalk was 90 degrees and the length of fruit stalk was 30mm and 35mm. In a single sweet pepper picking test, the image recognition time was about 5.3 seconds and the machine arm was activated for about 14.3 seconds. Complete an average of about 19.6 seconds for the entire picking trip. The integration test was conducted in the laboratory. Harvest range was 1500 × 1500mm, each time the visual shooting range of 500 × 500mm, the vehicle in accordance with the order of movement of harvesting test. The results of picking 58 sweet pepper model tests showed that the success rate was 93.7%. In the sweet pepper fruit picking test, a total of 42 failed 8, the success rate of 81%. The major reasons for the failure include the foam material of the gripper was too thin cause the folder failed to drop two sweet peppers, a sweet pepper cut fruit handles mistakes and five skew sweet pepper stalk cannot be effectively positioned to the stalk position. The vehicle takes 24.5 seconds on a single forward (500mm). The platform takes 6.3 seconds to lift up and 5.2 seconds for the lift down. When the visual identification area has a sweet pepper, the harvest time is about 21.8 seconds. The picking robotic system has been successfully developed and it will need to be further tested for sweet pepper picking operations in the greenhouse in the future.